copyright ©2016 by the fraser institute. all rights ... · on june 7, 2016, the government formed...

Fraser Institute d www.fraserinstitute.org

Copyright ©2016 by the Fraser Institute. All rights reserved. No part of this book may be reproduced in any manner whatsoever without written permis-sion except in the case of brief quotations embodied in critical articles and reviews.

As the researchers have worked independently, the views and conclusions expressed in this paper do not necessarily reflect those of the Board of Direc-tors of the Fraser Institute, the staff, or supporters.

Date of issue: October 2016.

Printed and bound in Canada.

Cover design and artwork: Bart Allen

Counting Votes: Essays on Electoral Reform

Lydia Miljan, editor

FRASERINST I TUTE

www.fraserinstitute.org d Fraser Institute

Contents

Preface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . i

Chapter 1: First-Past-the-Post: Empowered Voters, Accountable Government John Pepall . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

Chapter 2: Electoral Rules and Fiscal Policy Outcomes Jason Clemens, Taylor Jackson, Steve LaFleur, and Joel Emes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21

Chapter 3: The Consequences of the Alternative Vote Lydia Miljan and Taylor Jackson . . . . . . . . . . . . . . . . . . . . 45

Chapter 4: The Imperative of a Referendum Patrice Dutil . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81

Publishing information . . . . . . . . . . . . . . . . . . . . 104

Purpose, funding, and independence . . . . . . . 105

About the Fraser Institute . . . . . . . . . . . . . . . . . . 105

Editorial Advisory Board . . . . . . . . . . . . . . . . . . . 106

www.fraserinstitute.org d Fraser Institute d i

Preface

In its December Speech from the Throne, Justin Trudeau’s government reaffirmed the Liberal party’s campaign promise that 2015 would be the last election held under first-past-the-post (FPTP). While the government has no preference as to what should replace the 150-year-old FPTP system, it has ruled out holding a referendum to help guide the choice.

On June 7, 2016, the government formed the Parliamentary Commit-tee on Electoral Reform. The committee’s task is to deliver alternatives to our current electoral system by December 1, 2016. In the committee’s mandate, and in its five guiding principles, there is no mention of assess-ing the current system. While the government claims that “the principles do not prejudge an outcome,” the government’s pledge does preclude the committee from recommending the status quo. The authors of this volume are not restricted in this manner. Instead, they assess and evaluate FPTP against other systems to determine whether a different system could pro-vide a better arrangement for the Canadian electorate.

The parliamentary committee began its deliberations formally at the start of July by launching a website devoted to educating Canadians about electoral reform and encouraging them to engage in a “national dialogue.” Over the summer months the committee held over 20 public meetings and heard from electoral reform advocates, academics, and individuals. De-spite their diligence, a September 2016 Ipsos poll found that 81 percent Canadians were not aware that the committee had begun its work. It is


ii d Counting Votes: Essays on Electoral Reform

therefore not surprising that 55 percent of Canadian felt that there should be a national referendum to decide how future elections should be run in Canada.

The fact that so few Canadians are aware of the committee’s activities should not come as surprise; it is likely that many were unaware that the Liberal party made the campaign promise in the first place. Although the 2015 election campaign will be remembered as the longest one in over a century, electoral reform was not one of its dominant issues. In fact, it was only mentioned in passing during the first all-party debate in August, and ignored in all subsequent debates. While the Liberals promised that they would get rid of FPTP in their campaign platform, they provided few de-tails about why they made that commitment, or what they would replace the system with. The NDP and the Green parties had more concrete plans, each promising to replace the existing system with some form of propor-tional representation. Even the Conservatives made a passing reference to electoral reform by committing to holding a referendum before making any changes.

The media paid scant attention to electoral reform during the 11-week campaign, focusing instead on Syrian refugees, deficit issues, and follow-ing the leaders’ tours. It was only after the December throne speech that newspaper columnists began writing about the different electoral systems and making the case for one over another.

In this volume, we take a look at the issue of electoral reform without pre-judging any one system over another. John Pepall starts us off with his chapter, “First-Past-the-Post: Empowered Voters, Accountable Govern-ment.” His sober examination of why voting matters and the consequences of choosing an electoral system should give Canadians reason to pause and consider why we might want to engage in this process. At its core, Pepall argues that changing the system may affect the outcome of elections, but it can’t negate the fact that in a multi-party pluralistic system, there will be winners and losers. He aptly remarks that systems that allow voters to make more than one choice creates a paradox in that voters can avoid mak-ing discrete choices. In the end, there is a greater chance of minority and coalition governments.


Preface d iii

Some have argued that minority and coalition governments are harm-less; indeed, they often make the case that these types of government create more cooperation and collegiality. Jason Clemens and his co-authors ad-dress this issue in their paper, “Electoral Rules and Fiscal Policy Outcomes.” Their analysis focuses on proportional representation electoral systems and their propensity to incur higher debts and deficits than majority/plu-rality systems. The experience from major Western democracies is that systems with proportional representation tend to have more political par-ties than those with FPTP. Canada is an outlier for majority/plurality sys-tems with its five parties. But that does not mean that a PR system would reduce the number of parties; indeed, it is more likely that the number of parties would increase because of incentives in PR systems. Moreover, this increase in the number parties is directly related to an increase in coalition governments. From 2000 to 2014, countries with coalition governments had higher government spending and higher rates of borrowing than those with majority governments.

While proportional representation has more of a propensity to increase government spending as a percentage of GDP than our current system, what about other electoral systems, such as the alternative vote? The al-ternative vote (AV) system is similar to FPTP in that it is part of the family of voting systems called majority/plurality. It is also the only system that Prime Minister Justin Trudeau noted he had a personal preference for. This system offers an easy way to change the electoral system without changing the number of seats in the legislature—and it only changes the selection on the ballot. It is also associated with a decrease in strategic voting because voters rank order their choices. This means that rather than marking an X for the one party a voter wishes to be elected, voters rank their choices. If their first choice happens to have the fewest number of votes, but there is no clear winner, then their second and subsequent choices are taken into consideration. While this system may seem relatively easy to administer from the government’s perspective, it can have consequences in election outcomes. In the essay, “Consequences of the Alternative Vote,” Taylor Jackson and I examine poll data from the 1997 to 2015 elections to see if the outcome would be any different in an AV system than under FPTP. We


iv d Counting Votes: Essays on Electoral Reform

found that in all cases, AV benefitted the Liberal Party the most, and the NDP the second. In all cases, conservative parties lost seats. While we ac-knowledge that each election is unique and that election results from any one campaign have far-reaching consequences, we note that this system would require right-leaning parties to adjust their election strategy and policy choices more than the centre or left-leaning parties. Moreover, the study also suggests that in the last 23 years, an AV system would have been unlikely to change the outcome of elections where Liberals were the victor under our current system. But it would have changed the outcome for the minority Conservative governments—and their 2011 victory.

While the current government’s five guiding principles are laudable in that they seek to produce a fairer electoral system, one of the most glaring omissions in those principles is that they place no value on a change in the governing party. According to Freedom House, one of the indicators of a free and fair election is that there is “a realistic opportunity for the oppo-sition to increase its support or gain power through elections.” A central tenet of electoral success is that there is a fair opportunity for other politi-cal parties to have a realistic chance at changing the current regime. This is not only important for political parties themselves, but it is crucial for the renewal of countries to ensure that they have policy changes. The AV sys-tem fails to address any of the five values the government seeks to address in its electoral reform initiative, and in fact, it would create a new problem: our elections would be less competitive in the future.

This leads to the question of the role the public should have in the elec-toral reform process. The all-party committee has put a significant amount of effort into inviting the public to be engaged in the discussion. Their web-site encourages individuals to make a submission, attend an event, and host a dialogue. They even provide a helpful guide on how one might host such an endeavour. Yet, as noted above, Canadians seem not only unaware that this process has commenced, but that it is nearing its end. Patrice Dutil, in his chapter, “The Imperative of a Referendum,” argues that not having a referendum on this issue could be considered unconstitutional. He points out that when the provincial governments of Prince Edward Is-land (2005), British Columbia (2005 and 2009), and Ontario (2007) sought


Preface d v

to change their electoral systems, they each engaged the public through a referendum. He also argues that parliamentary convention dictates that a referendum be held. Dutil points out that the Supreme Court has laid out a series of questions to test the validity of a convention. The so-called “Jen-nings Test” provides three conditions that have to be met in order for there to be a constitutional convention. Dutil makes the case that all three are satisfied in the issue of a referendum on electoral reform.

The essays in this book were chosen to provide the reader with a nu-anced overview of the current electoral system and the potential conse-quences of choosing a different system. While the government seeks to increase voter engagement and attain more representatives from minor-ity groups, it is growing clear that changing the electoral system may not achieve any of those goals. Moreover, as John Pepall reminds us, there is no perfect electoral system. Each system has its own benefits and drawbacks. The issue facing the parliamentary committee is whether a new electoral system can change for the better the political culture of a country that has enjoyed free and fair elections for the past 150 years.

—Lydia Miljan, editor

www.fraserinstitute.org d Fraser Institute d 1

CHAPTER 1 First-Past-the-Post: Empowered Voters, Accountable GovernmentJohn Pepall1

Introduction

In its election platform, and again in its December 2015 Speech from the Throne, the federal Liberal government committed itself “to ensuring that 2015 will be the last federal election conducted under the first-past-the-post voting system.” The implication is that our present way of voting, more technically and politely called single member plurality voting (SMPV) is deeply flawed.2 The government does not address what the flaws of first-

1 John Pepall is the author of Against Reform published in 2010 by the University of Toronto Press. He took degrees in philosophy and politics at Trent University and law at York University. He was called to the Ontario bar in 1978 and practiced civil litigation in Toronto. He was a Progressive Conservative candidate in the Ontario election of 1990. He has written on politics, history, the law and the arts for The Idler, The Literary Review of Canada, The Ottawa Citizen, The Times Literary Supplement, The Dorchester Review, The National Post, and other publications. He contributed a chapter to Rethinking the Constitution published by the Oxford University Press in 1996. His paper, Laying the Ghost of Electoral Reform, was published by the Macdonald-Laurier Institute in 2011.

2 While single member plurality voting is technically the more correct term for this form of voting, for the sake of clarity and consistency with the other papers in this book, we will be using the more widely used “first-past-the-post” (FPTP).


2 d Counting Votes: Essays on Electoral Reform

past-the-post (FPTP) are and how another way of voting could overcome them.

There is to be a “national engagement process” to decide on electoral reforms, but the argument that FPTP is the best way of electing a Parlia-ment has been pre-emptively excluded from the discussions. As FPTP is the best way of electing a Parliament, this is unfortunate.

For reformers it is obvious that FPTP is bad. But, if they are right, the result of practically every election in our history was wrong and most peo-ple don’t seem to care. Reformers are heartened by polls that show major-ity support for reform, but in referendums voters have repeatedly rejected specific reforms. Most people don’t seem to care about what, for reformers, must be the most important issue in politics. While reformers think the case for reform is obvious, some also think that the issue is too compli-cated to be addressed in a referendum. Of the people who tell pollsters that they favour reform, what they want from elections is what we get now and would not get from any reform.

Electoral reform can get very complicated. But the complications are the fault of the reformers, who do not understand the purpose of voting and who are trying to do the impossible.

Whatever the government may be thinking, once Pandora’s Box has been opened, there is no telling where the discussion will go. The NDP is one party that knows what it wants. Others will weigh in. All likely options must be considered—and rejected.

Decisions, pluralities, and preferences

At the root of all arguments for electoral reform is the complaint that can-didates can win seats and parties form governments with only a plurality of the votes.

“Majority rule” is widely understood as the basis of democracy. It is ob-vious that when there are more than two choices, there may be no majority for any of them. It is not as obvious, but just as true, that there is no way of


First-Past-the-Post: Empowered Voters, Accountable Government d 3

contriving a majority by adopting some other way of voting from the one that occurred naturally to people when they first decided things by voting.

Voting is a way of making decisions. When a group of people is having difficulty deciding what to do, someone is likely to say, “Let’s put it to a vote.” That usually means a yes or a no and a majority decides.

But if friends trying to decide between three restaurants to go to found there was no majority for one, they might be tempted to list their prefer-ences, and after eliminating the one that got the lowest preferences think they had found a majority when there previously hadn’t been one.

Like many people keen on electoral reform, they haven’t heard of the “paradox of voting.” It can be illustrated simply in an example when three voters [1, 2, and 3] have to choose between three options or candidates [A, B, and C] by a preferential ballot:

1 2 3 A B C B C A C A B

What do we find? A majority (1 and 3) prefer A to B. A majority (1 and 2) prefer B to C. So we can eliminate C and have a runoff between A and B? No. Because a majority (2 and 3) prefer C to A.

If an individual said he preferred A to B, and B to C, and C to A, we would say that he could not make up his mind. But this paradox lurks in all voting where there are more than two choices, in all the schemes intended to improve our elections. There is no problem when there are only two choices: A majority will select one, and that majority will be accepted. But where there are more than two choices, electoral schemes that appear to produce a majority by a process of elimination ignore the paradox of voting.

It gets worse. Proponents of electoral reform may say, “What does it matter? There is a result and surely it is better than someone getting elected with only 30% of the votes.” But preferential voting’s orderliness conceals a fatal flaw. It can breach monotonicity, the core principle that a candidate should be better off she gets more votes. This is not, as it surely should be, necessarily so.



Suppose there are three candidates and 21 voters with the following preferences:

1st 2nd 3rd7 A B C3 B A C5 B C A6 C A B

On first preferences, B gets 8 votes, A gets 7, and C 6. With the lowest number of votes, C is eliminated. Because of the number of those who had C as their first preference and had A as their second, on the second count, A wins:

13 A8 B

Suppose the three voters whose preference was B A C had switched to A B C:

1st 2nd 3rd7 A B C3 A B C5 B C A6 C A B

Now B has the lowest number of votes on the first count and is elimi-nated. Because the B voters’ second choice is C, on the second count A loses and C wins:

10 A11 C

More votes for A, and no more votes for C, has nonetheless led to A losing and C winning. It can also happen that a shift of votes between two candidates makes no difference to them but may cause one or more other candidates, whose votes remain the same, to win or lose.



In 1951, the economist Kenneth Arrow published his General Impos-sibility Theorem, which proved that “there is no method of voting which will remove the paradox of voting… no matter how complicated” (Arrow, 1951: 59). It was cited as a basis for his Nobel Prize. The paradox of voting and the breach of monotonicity from preferential voting do not mean that voting is pointless. They only point the way to understanding voting and making it work to enable voters to decide.

One complaint made against FPTP is that it may force people to vote “tactically,” in other words, voters may not vote for the candidate they “sincerely” want to win, but for another with a better chance of winning. Reformers claim that in a different system than FPTP, voters will be able to vote “sincerely,” not worrying that their votes may let in someone they are against on a plurality. But with preferential voting, voters cannot even know whether their vote will help or hinder their candidate.

And while many voters may have preferences running down through three or more choices in an election, many may not. They are for one can-didate or party and don’t like any of the others. Reformers like to complain that 69% of voters voted against a candidate who wins with 31% of the votes. But the 69% did not vote against the winner. They voted for some-one else. If voting against the winner was their intention, they could have coalesced around the runner-up. Those who advocate preferential voting implicitly concede that under plurality voting voters cannot be presumed to vote against anyone.

Media reports suggest that the government leans to preferential voting as the alternative to FPTP (Clark, 2015, August 7). Analysis shows that had preferential voting been in place in Canada, based on the voting patterns in the October 19, 2015 election, an even greater Liberal majority might have been produced than the one the Liberals won then (Jansen, 2015). The insinuation is that the electoral reform commitment was just a Liberal scheme to ensure it receives big majorities in perpetuity.

But it is a mistake to assume the outcome of any electoral reform based on existing voting patterns. Faced with a different ballot, voters may vote differently. And, in any event, voting patterns and party standings will evolve as issues, interests, and ideas change over time. Whatever political



strategists may calculate, or disinterested commentators speculate as to the advantages or disadvantages of preferential voting in particular cir-cumstances, those calculations have to be judged based on abstractions from existing voting patterns.

Preferential voting attempts the impossible: making a majority where there is only a plurality. Its proponents claim it is simple. All you have to do is mark 1, 2, 3…. But, if filling out a preferential ballot is only a little more complicated than voting as we do now, what happens to the preferences when they are counted is beyond the understanding of voters.

The long term consequences of adopting preferential voting in Canada are unforeseeable. It might well, for some time, entrench in power a “natu-ral governing party.” Or, with party splits, new parties emerging, regional tensions, and the evolution of ideas, interests, and personal rivalries, lead to fragmentation that is beyond voters’ ability to control.

Preferential voting in Australia may offer a lesson. For generations and for all intents and purposes, Australia had a two-party system that alter-nated in government: the Labour Party and the conservatives. The latter was made up of a permanent coalition between the Liberal and Nation-al parties. Preferential voting allowed the parties in the coalition to sort things out between themselves. Late in the last century things began to fall apart. Independents and fringe parties took seats and sometimes held an unstable balance of power.

Where there is no majority, a plurality is just as valid a basis for a group decision as a majority. It is the most popular, it is transparent, both voters and candidates can see how it works and govern themselves accordingly, and it is a decision. Some may be unhappy with the result. But so may 49.99% when a majority decides.

Proportional representation

While preferential voting claims to solve the alleged problem of the MP for Maple River being elected with only 31% of the votes, it does nothing to produce proportional representation (PR) of parties, to solve another



alleged problem—that of the Liberals winning a majority (over 54% of the seats in the House of Commons in the most recent federal election) with under 40% of the votes.

That FPTP does not produce proportional representation of parties is obvious, though advocates of PR never tire of giving examples of its failure to do so. The question of why we should want proportional representation of parties is never addressed. It is assumed that we should, that PR is “fair.”

To understand PR, and why it is bad, we need to ask what parties are for.Parties are a paradox. The root of the word “party” means division. But

parties serve a necessary purpose–not by dividing people, but by bringing them together.

Parliament had a long history before there were parties. MPs in Eng-land’s House of Commons in the 16th century may have seen things dif-ferently depending on whether they represented London, or a town, or country seat. But there were no parties.

It was only with the coming of what is now known as “responsible gov-ernment” in the early 18th century that parties finally emerged. As respon-sibility for government came to be assumed by a ministry drawn from Par-liament and supported by the House of Commons, it became necessary to organize that support and seek its continuance in elections. Nothing at all like modern party organization emerged until well into the 19th century. But in Parliament, and at elections, parties finally emerged as the purpose for which they exist emerged.

A party exists to form a government. It is a political association of peo-ple whose interests and ideas and confidence in each other make it possible that they should be able to work together to support coherent measures, a ministry, and, most important, a budget. No party has any value in politics unless it is a potential government.

A party exists to form a government. This is not a dogma. It is an his-toric fact. If it is not true, what is a party for? It is not enough to say that it exists to promote interests or ideas. There are all kinds of associations that promote the interests and ideas that are the stuff of politics. But to be a political party an association must put up candidates selected by its own rules and run some kind of campaign on their behalf. Any association that



does that assumes a responsibility for all that governments may do, even if the protection of the environment or the independence of Quebec or farmers’ interests are what brought them into politics, and even if their position on everything else is indifference or that nothing should be done. That is a position in itself and the one-issue or one-interest party will have to contend not only with those who don’t care so much about farming or whatever, but with those who insist that other issues have to be addressed and other interests weighed. MPs are called upon to deal with all our pub-lic business. They cannot choose to deal with only some of it. Nor can vot-ers in choosing an MP limit themselves to some issues and interests. The government they elect cannot and will not.

Under PR small parties with no ambition to form a government by themselves win seats and often join coalition governments. So they may claim they advance the ideas and interests they stand for. But what is really happening? Either their ideas and interests are supported by most voters, in which case they should not be small parties but big parties, or their ideas and interests are shared with other parties and their claim to be the sole champion of those ideas is false. Conversely, their ideas and interests are supported only by their 10% of the voters and they lever their 10% of Parliament to impose those ideas against the wishes of most voters. Or they abandon their ideas to take places in government and so betray their supporters. The only basis on which small parties can claim to advance their interests and ideas without accepting the challenge of getting a plu-rality of votes is by seeking disproportionate power in a fragmented Parlia-ment. Hence the fondness of small parties for PR.

WasteA favourite claim of PR’s advocates is that votes in our elections for losing candidates and excess votes for winning candidates are “wasted.” Most of those votes are cast by voters who know that their candidate is either sure to win or sure to lose, but in the face of all the carping, voters in their mil-lions cast their votes anyway. They see the point of doing so even when the advocates of PR refuse to. Advocates of PR could not come up with the fig-



ures showing what they claim are wasted votes, if voters were not so keen, as they see it, to waste their votes.

The concept of the wasted vote is an artefact of PR doctrine. You could say you wasted your vote if you did not vote, in the same way as if you had a ticket to a ball game and did not go you would say you wasted your ticket. Or you could say you wasted your vote if you spoilt your ballot. But to say you have wasted your vote if your vote does not count, in however minuscule a degree, towards a party winning a seat, amounts to saying that your vote is wasted if it does not produce a winner. Under proportional representation, in Canada every vote would count for 0.00002 seats. But the price of this tiny victory would be loss of control over government. It is only when you accept the proportional representation argument that every vote must count towards seats that the phrase “wasted vote” takes on meaning. When you vote for Jones for mayor, and Singh wins, or Jones wins by a landslide, you do not think you wasted your vote.

Voting is a procedure for letting the people decide. If they are not unan-imous, any decision must involve one choice winning and one or more losing. The PR goal that “everyone wins and all get prizes” means that the voters do not decide. That is a real waste of voting.

Two schemesThe simple and logical way to achieve PR is for voters to vote for a party list of candidates in order from 1 to whatever number of seats are open. If a party gets 10% of the vote, enough candidates at the top of the list to fill 10% of the seats would be elected. In Canada there would probably have to be separate lists for each province and territory to comply with the Consti-tution, producing ragged results across the country. The Territories, with one seat each, would still have first-past-the-post.

Simple party list PR has never been strongly promoted in Canada, per-haps because, despite their willingness to endorse the principle of PR in polls, voters are leery of parties, and reformers realize that asking voters simply to vote for a party would not be popular.



Two elaborate alternative schemes were put to voters in three prov-inces early this century and roundly rejected. Each illustrates the folly of trying to do better than best.

Single Transferable Vote (STV)

The Single Transferable Vote was proposed by the British Columbia Citi-zen’s Assembly in 2004. It came closer than any other electoral reform ini-tiative to being adopted in Canada. It is seductive in its elaborateness. The very highest claims have been made for it. Keen proponents of electoral reform have called STV a cult (Sykes, 1990: xii).

It works like this: Ridings would elect more than one member. We shall use the hypothetical riding of Fraser as an example. In that riding, three members are to be elected. Voters would be asked to mark the order in which they prefer the candidates on their ballot. The number of ballots is divided by the number of members to be elected plus one. The next full number above that is the “Droop quota,” the number of votes needed to elect a member. So if there are 10,000 ballots, a candidate needs 2,501 votes to be elected.

If a candidate gets 2,501 or more first preference votes he is declared elected. If he gets more than 2,501 first preference votes--say 3,000--then a transfer value is calculated for the second preferences on the ballots on which he was the first preference by dividing the surplus votes by his total number of first preferences. So these second preferences are transferred at a value of 499/3000.

If no candidate gets 2,501 votes, the candidate with the fewest votes is eliminated and the second preferences on the ballots on which she was the first preference are counted. The process of transferring votes goes on until three candidates have got 2,501 votes and are declared elected. Votes for eliminated candidates can be transferred several times at full value if they do not land with a winning candidate and surplus votes for winning candi-dates can be transferred at ever smaller fractions. The elaborate arithmetic of the single transferable vote encourages the hope that it makes every vote



count even when all but a handful of experts have lost track of what is hap-pening to the votes.

Why are there to be three members elected in the riding of Fraser? STV needs multi-member ridings because it is designed to allow the election of members by a small percentage of votes. The bigger the riding, the smaller the Droop quota and the percentage of votes needed to elect a member. While the percentage of voters needed to elect a member declines, the number of members elected to represent voters who do not support them, even vehemently oppose them, rises.

The riding of Fraser elects three members, but an individual voter is only, at best, electing one. Once a voter’s preferences have “counted” to-wards the election of one member, they are cast aside. Any other member elected in the riding no more represents that voter than the member for Peterborough represents a voter in Medicine Hat. It is not a question of whether the voter likes or would have voted for that other member. He could not, no more than a voter in Peterborough can vote for the member for Medicine Hat. He may have placed the other member somewhere on his list of preferences, but that preference would not have been counted.

All the candidates’ names are on the ballot, but the voter does not get to choose three candidates to fill the three seats. If the first preference can-didate, the one with the real vote, is elected, the ballot may be cast aside. Conversely, it is possible to fill in all preferences so that literally the last person you would want elected wins your vote.

STV appears to seek the voter’s will in exhaustive detail, but in actual fact it restricts the voter’s say to what STV deems a “fair share.” STV may appear to ask which three candidates you would like to see elected, but it is really only asking which one of them you want.

STV is strongly sold as a means of achieving proportional representa-tion, but claims to offer the best of both worlds by asking voters to choose between individuals. Independents with no party affiliation can be elected. In fact what it really does is allow the election of members who have se-cured only a small fraction of the vote. In a riding with six members, just over 14% of the vote would be enough for each one to be elected. In our present elections, some voters complain that members are elected with as



little as 30% of the vote. If that is the case now, how can a member who got less than 15% of the vote speak for her riding?

While barring the voter from effectively choosing all the members who are to represent her, STV asks for a range of choices that is unrealistic. A preference of the Liberal voter between the New Democrats and the Conservatives and then between the individual New Democrat and Con-servative candidates may be expressed but cannot be serious. It may be mischievous. Voters evidently have some difficulty deciding how to cast a single vote. Their choice of multiple preferences of decreasing value must become meaningless. A preferential ballot asks for far more choice than voters can seriously make. And any vote may become a vote for a candidate at full value.

STV’s elaborateness suggests that it captures the voter’s will with great precision, but it can only do so if the voter’s will is precisely cut to STV’s theory. Take, as an example, voting in a three-member riding. Mary lists the three Suede Party candidates: 1, 2, and 3. STV registers a preference in that order. Very likely Mary has no preference between them—at least none between 1 and 2, or 2 and 3. The preference STV registers is an arte-fact. But it may be enough to elect 1 before the preferences of others who strongly prefer 2 over 1 come into play. Mary did not mean to give 1 an advantage over 2, but she did. Or 3 may be the one Suede candidate with a chance, but be eliminated before Mary’s 3rd preference is counted. As the counting goes on, things become even trickier. Mary votes “sincerely,” as STV wants her to do, that she prefers 4 to all the others left in the race. But when Mary’s 4th preference is counted, the real contest is between her 6th and 7th choices, and the 7th is elected while 4 is eliminated.

STV claims by its contrivance and mathematics more effectively to ex-press the will of the voters than any other system. In the end, all that can be said is that STV produces the result that STV produces.

STV is a form of preferential voting. As explained above, preferential voting breaches monotonicity. Candidates can win or lose because they got fewer votes or more votes, or some other candidate got more votes or fewer votes. Voters haven’t a clue what the consequences of their prefer-



ences may be, though apparently in Australia fringe parties have worked out ways of gaming the system to get senators elected on .5% of the vote.

Given that it can be shown that STV gets the will of the voters wrong and that more votes can lose a candidate an election, STV’s claims are demonstrably false.

Mixed member proportional voting (MMP)

The proposal before Ontario voters in October 2007, and PEI voters in No-vember 2005, and favoured by the NDP, was so-called Mixed Member Pro-portional voting (MMP). It is used most notably in Germany, but also in New Zealand since 1996. As proposed for Ontario, it would have worked like this: 90 members of the provincial Parliament would have been elected in the old-fashioned way in single member constituencies. A further 39 members would have been elected from party lists on the basis of propor-tional representation. Many may think that this is a kind of compromise, with 39 of 129 members being elected through proportional representa-tion—not even half PR. Proponents of the scheme are content that they think that. But under MMP, the list seats are not assigned in proportion to party votes. They are used to see that, as far as is mathematically possible, each party gets seats in proportion to its party vote. So assume there is a party none of whose candidates won one of the constituency seats, but which got 10% of the party vote. It would get 10% of the seats. Conversely, assume there is party that received 45% of the constituency vote and got 45% or more of all the seats, easy enough to achieve as we are continually reminded by PR’s fans. It would get none of the list seats. The list seats would not “introduce an element of proportionality,” in the phrase used by the Law Commission’s report on electoral reform from 2004; those list seats would impose proportional representation.

The 2007 report of the Ontario Citizens’ Assembly had as its title One Ballot, Two Votes, ostensibly offering the promise of more choice. But the voter does not get more choice. Each ballot asks for a vote for a constitu-ency candidate and a vote for a party. But the voter who votes for a win-



ning candidate effectively cancels her second party list vote. Only the voter whose constituency vote is, in the theory of proportional representation “wasted,” can cast an effective list vote.

The only “mixed” element in MMP is the retention of single-member constituencies. MMP’s proponents assure us that we should still have our local member for a riding only a bit larger than it is now. Half or fewer of the members would come from party lists and not represent a riding. While the retention of local members is a good thing, so far as it goes, even local members under MMP would necessarily be more party creatures, as their local success or failure would redound on the party list.

The belief that retaining local members while rigorously imposing PR strikes a good balance results from an overly abstract analysis of voting in which our vote is broken down into a vote for a local representative and a vote for a government, a vote for a person and a vote for a party. When we vote now we are not torn as academics suppose between these two or any other factors. We weigh them and much more in making a decision. Is Smith in favour of gun control, lower taxes, bombing ISIS, a good speaker but an indifferent servant of the constituents, famously accomplished but a stranger to the riding?

We cannot split our votes for each consideration: a vote for someone to speak in the House, a vote for someone to serve the constituents, a vote on this issue and that. It is not simply impractical but wrong, because, what-ever the considerations of policy and principle, ability and character we may weigh, we are voting for a human being to represent us. And that human being will be called upon to do all that a member may do, and to address all the issues that may arise.

Though independents can be elected in the constituency seats under MMP, most constituency members are elected for a party, and the party is the prime consideration in voters’ minds. But supporting a party in an election now means hoping that it will form a government. Under MMP there would be little hope of one party forming a government.

MMP would produce two classes of MP. Constituency members would have their constituency duties with the work and hope of local strength



that they offer. List members would be free of local duties but indentured party servants.

Advocates of MMP tout this system as offering more choice. But as we have seen, this is not true for the voter who votes for X party’s candidate in a riding and X party’s list. MMP’s advocates proclaim that one can split one’s vote: vote for X party’s candidate and Y party’s list. But why would one want to do that? Suppose Lee is the candidate of the X party. You vote for Lee because you think Lee is bright, honest, hard-working, and so on. If Lee were running for the Y or Z party, you would still vote for Lee. As people used to say, you “vote the man, not the party.” If that is how you see things, then party cannot mean much to you, and your party vote cannot mean much to you either.

MMP fans think of splitting your vote as being broadminded, being not bound to a party. But as proportional representation, which is what MMP is, is designed entirely to serve parties and entrench them, its advocates cannot claim indifference to party.

If you decide to split your vote, the likelihood is that you have figured out how MMP really works and know your list vote may not count because your party already has as many seats as it is allowed under PR. So you vote “tactically” for the party most likely to ally itself with your real party choice. A risky vote, as post-election alliances can be unpredictable. Ideally, you could vote to elect a candidate of one party in your riding and support the list of another party that would agree on everything with your riding candidate’s party.

Thus is born the idea of a decoy party. Suppose there is a 200-seat house with 100 seats filled by riding candidates and 100 from party lists. There is a Tory Party and a Conservative Party. Tories run in the ridings but not Conservatives. With 35% of the vote, the Tories elect 45 members in the ridings. There is no Tory list for the party vote but there is a Conservative list. It gets 35% of the vote and 70 list seats. The Tory/Conservatives have elected 115 members, 57.5%, with 35% of the vote. Hard to do even with our present way of voting.

Blatant decoy parties could be banned by legislation. But the possibility of decoy parties shows a fundamental flaw in MMP. The two votes on one



ballot are only useful or effective if there is something like a decoy party available. In practice, parties can become effective decoys without being so bold in their planning that they will be caught by legislation. In Germany, the Free Democrats operated as a decoy for the Christian Democrats. The differences between them on policy were no greater than differences with-in the Christian Democrats and the relations between their politicians as cordial or bitter as those amongst Christian Democrats. In election after election, Christian Democrat voters having elected a constituency mem-ber gave their list vote to the Free Democrats, accounting for roughly half their votes.

The decoy problem is peculiar to MMP, but typical of the problems that arise with every contrivance to do better than the system we have now. Ev-ery complication intended to right imagined wrongs produces real wrongs.

Making the sorry best of MMP depends on a thorough grasp of how it works and a knowledge of what one’s fellow voters will do that few polls can give. There is every reason to believe that many people will not un-derstand how it works and what they should do. Surveys done in Scotland after MMP was adopted for elections to the Scottish Parliament showed woeful incomprehension. Under the slogan “Second vote Green” the Scot-tish Greens played on this, encouraging voters to see the list vote as a sec-ond preference, when, of course, for the overall result the list vote rules (Commission on Boundary Differences and Voting Systems [“Arbuthnott Commission”], 2006: 31, paras 412 et seq.; see also Herbert, Burnside, Ear-le, Edwards, Foley, and Mciver, 2007: 42 and 46).3

Voters under MMP will fall into two groups. The majority, probably, who don’t understand how it works, and the rest looking for a decoy.

MMP has all the faults of proportional representation. What distin-guishes it (the retention of locally elected members), serves only to ob-

3 The Arbuthnott Commission said, perhaps rightly, that the Greens were not to be blamed for campaigning to seek their best advantage from the voting system. But the system must be blamed if the Greens got half or more of their seats from voters who did not understand the consequences of their second vote, as appears to have happened. A change in the form of the ballot paper and voters understanding the system better perhaps led to the Greens falling from seven seats in 2003 to two in 2007.



scure its faults and compound them by confusing voters or opening the way to decoy voting by the astute.

If deciding is taken away from voters

What coalitions are formed after an election under PR or preferential voting, and how parties to a coalition work together, are beyond the control and even the understanding of voters. Coalitions themselves are the product of an attempt to make a majority where there is none and a decision where there has been none. Social choice and game theory might be invoked to explain what happens, but even the participants do not understand what happens. Voters may think that voting Green will make a greener govern-ment, or voting Conservative will make a more conservative government, but they may be wrong.

Electoral reformers idealize coalitions as benign compromise while complaining of the tendency of parties under plurality voting, or first-past-the-post, to seek the effective centre. Under plurality voting, parties seek the most votes by offering a program with the widest appeal and stand or fall on it. Under proportional representation, parties offer a program for a niche with no way of knowing, or assuring voters, that they can be effective.

The failure of elections under proportional representation to produce decisions leads to unstable or stagnant governments and can shift power from the politicians and those who choose them to the bureaucracy. We need no Marx, or Djilas (author of The New Class), to see that bureau-cracy is a class with its own interests in big government and control. If voters want that, they can have it—as they have in many countries without proportional representation. However, the entrenchment in power of what is, from one perspective, an interest group, without strong direction and control by politicians and voters, must always threaten freedom and ac-countability.

It is the ability to “throw the bums out,” more even than the ability to choose a new government, that is the most striking practical virtue of FPTP. Our governments are responsible, must answer to the voters, and



are regularly defeated. Joseph Schumpeter (1987: 272) and Karl Popper (1963 and 1988, April 23) saw the ability to get rid of an unsatisfactory gov-ernment as the purpose and test of democracy and condemned propor-tional representation for not seeing this. To “throw the bums out” is almost impossible with proportional representation. In the 50 years after 1945 in 103 elections in Belgium, Germany, Italy, Japan, the Netherlands, Sweden, and Switzerland, the major governing party was only thrown from office six times (Pinto-Duschinsky, 1998, September 25). Major parties have re-mained in government for decades under proportional representation de-spite wide fluctuations in their votes. Minor parties often seem to share in government in inverse proportion to their electoral success, turfed out when their vote grows and they look threatening, and brought in when it sags.

Conclusion

That plurality voting is old is nothing against it. It is still the most widely used way of voting. The plethora of alternatives, only touched on in this paper, shows that no better way has been found, though the alternatives were contrived and tried from the 19th century and have served the inter-ests of a political class, generally state subsidized.

First-past-the-post is the only way of voting that assures effective and accountable government. We abandon it at our peril.

References

Arrow, Kenneth J. (1951). Social Choice and Individual Values. Yale University Press.

Clark, Campbell (2015, August 7). In reforming voting system, parties favour what benefits them most. Globe and Mail. <http://www.theglobeandmail.com/

news/politics/in-reforming-voting-system-parties-favour-what-benefits-them-

most/article25745735/>, as of July 5, 2016.



Commission on Boundary Differences and Voting Systems [Arbuthnott Commission] (2006). Putting Citizens First: Boundaries, Voting and Representation in Scotland. The Stationery Office Edinburgh.

Dummett, Michael (1971). Principles of Electoral Reform. Oxford University Press.

The Economist (1988, April 23). The Open Society and Its Enemies Revisited. The Economist.

Herbert, Stephen, Ross Burnside, Murray Earle, Tom Edwards, Tom Foley, Iain Mciver (2007, May 8). Election 2007. SPICe Briefing. Scottish Parliament Information Centre. <http://www.parliament.scot/SPICeResources/Research%20

briefings%20and%20fact%20sheets/SB07-21.pdf>, as of June 15, 2016.

Jansen, Harold (2015). Reforming the electoral system Policy Options. <http://

policyoptions.irpp.org/fr/magazines/november-2015/reforming-the-electoral-

system/>, as of July 5, 2016.

Law Commission of Canada (2004). Voting Counts: Electoral Reform for Canada. Law Commission of Canada.

Ontario Citizens’ Assembly on Electoral Reform (2007). One Ballot, Two Votes: A New Way to Vote in Ontario. Ontario Citizens’ Assembly on Electoral Reform. <http://www.citizensassembly.gov.on.ca/assets/One%20

Ballot,%20Two%20Votes.pdf>, as of July 5, 2016.

Pinto-Duschinsky, Michael (1998, September 25). Sending the Rascals Packing. The Times Literary Supplement.

Popper, Karl R. (1963). Chapter 7: The Principle of Leadership. The Open Society and its Enemies. Volume 1. Harper Torchbooks.

http://www.parliament.scot/SPICeResources/Research briefings and fact sheets/SB07-21.pdf

http://www.parliament.scot/SPICeResources/Research briefings and fact sheets/SB07-21.pdf



Popper, Karl R. (1988, April 23). The Open Society and Its Enemies Revisited. The Economist: 19-22.

Schumpeter, Joseph A. (1987). Chapter XXII: Another Theory of Democracy. Capitalism, Socialism and Democracy, 6th edition. Counterpoint Unwin Paperbacks.

Sykes, Leslie (1990). Proportional Representation: Which System? Hornbeam.


CHAPTER 2: Electoral Rules and Fiscal Policy OutcomesJason Clemens, Taylor Jackson, Steve LaFleur, and Joel Emes1

Introduction

The federal Liberal government’s commitment to reform the nation’s cur-rent electoral system in time for the next election has vaulted the issue of electoral reform to the forefront of public discourse (Liberal Party of Can-ada, 2015: 27). While there are a number of political and electoral aspects of this policy change to consider, one area that has been ignored thus far is how electoral rules influence government policy, particularly fiscal policy. Changing the method by which a society elects its political representatives changes the incentives and power structure of the government, which in turn influences fiscal policies.

The preponderance of empirical evidence clearly shows that moving towards a proportional electoral system would lead to higher govern-ment spending and more deficits. Put simply, changing our electoral rules doesn’t just change the way we elect our political representatives; it could also fundamentally change basic government policy over time.

The first part of this chapter provides some basic definitions and clari-fies several terms that will be used throughout the book. The second sec-tion provides an overview of the basic conceptual argument relating to

1 The authors are all analysts with the Fraser Institute. Please see the end of this chapter for their full biographies.



why electoral systems influence not only electoral outcomes but also fiscal policy. The third section reviews existing research on the relationship be-tween electoral rules and fiscal policy outcomes with particular emphasis on government spending and fiscal balance (deficits/surpluses). The chap-ter ends with a brief conclusion.

I. Terms and definitions

While covered in other chapters in this book, it is nonetheless important to define a number of key terms prior to reviewing the existing research on the relationship between electoral systems and fiscal policy. There are two broad areas of research within this sphere: governance and electoral rules. The former relates to the system of governance in place in a specific juris-diction. These systems range from parliamentary, such as exists in Canada where there is no formal or constitutional separation of power between the executive and legislative branches of the government, to presidential sys-tems, such as exist in the United States where the executive function (pres-ident) is constitutionally separated from the legislative powers (Congress).

There is no proposal to change Canada’s governance system, so the fo-cus of this chapter is on electoral rules.2 There are many different types of electoral systems, but they can be grouped into four basic types: 1) plural-ity/majoritarian, 2) proportional representation, 3) mixed systems, and 4) other systems.

Plurality/majoritarian systems select their representatives based on the largest share of votes won in specific electoral districts. The winning can-didate is not necessarily required to win a majority of the overall votes cast; a number of the plurality/majoritarian models rely on plurality voting, in other words, candidates are elected if they receive the highest percentage of votes in each electoral district (rather than an outright majority of all votes cast). Canada’s system of first-past-the-post (FPTP) is a plurality sys-

2 For an overview of existing electoral systems, see Reynolds, Reilly, and Ellis (2008), and O’Neal (1993).


Electoral Rules and Fiscal Policy Outcomes d 23

tem. It is formally referred to as a single-member plurality system. Major-ity systems are used when there is a desire to have the winning candidate receive an absolute majority of votes cast; such systems include the alter-native vote (AV) or two-round system (TRS).

The second category of electoral system is referred to as proportional representation (PR). These systems are designed to allocate elected repre-sentatives in a way that is based on the proportion of votes received. Such systems require multi-member districts so that elected seats reflect the broad proportion of the votes that different parties received. There are two main types of PR systems: party list systems and the single transferable vote (STV).3

Mixed systems use elements of plurality/majoritarian and PR systems. Other systems, meanwhile, are so diverse that they defy group classifica-tion. For the purposes of this essay, we focus on the differences between plurality/majoritarian and PR systems.

II. Why electoral systems affect fiscal policy: A conceptual and empirical argument

The link between electoral rules and fiscal policy is often over-simplified, which leads to a misunderstanding regarding the linkage. Indeed, the link

3 In addition to electoral formulas, economists Torsten Persson and Guido Tabellini also stress district magnitudes and ballot structures as influential aspects of electoral rules. “District magnitudes” is the number of legislators elected per voting district. This can range from a single-member district—as is currently the practice in Canada—to a single district for the whole country, such as is used in Israel, where all members of the Knesset are chosen by the entire electorate. “Ballot structures” refers to the specific manner in which citizens vote for their representatives. Ballots are structured in two common ways. Typically voters are either asked to vote for specific candidates (as they are in Canada), or to vote for a party list. While there are different ways that party lists can be structured, typically they are closed lists, meaning that the party determines the order of the candidates and then allocates the seats the party wins to members on the list in descending order. This means that if a party’s proportion of the vote share dictates that they receive nine seats, candidates one through nine on the list are elected, while candidates ten and below are excluded from the legislature.



between the two is based on an intermediary effect: the number of politi-cal parties elected.4 It is not a plurality/majoritarian or PR system per se that influences fiscal policy, but rather the incentives embedded within each that affect the number of political parties elected. In general terms, PR systems have a built-in incentive to elect more political parties, while plurality/majoritarian systems have incentives that lead to fewer parties.5

A substantial amount of research supports this general view. For in-stance, Persson and his colleagues examined 40 parliamentary democra-cies from 1960 to 1998 to assess the relationship between electoral rules and government spending. Part of their analysis looked at the difference between plurality/majoritarian and PR electoral systems, and whether they governed by coalition or majorities. They found that:

Compared to proportional rule, plurality/majoritarian rule is cor-related with a less fragmented party system, a lower incidence of coalition government, and a higher incidence of single-party govern-ments. (Persson, Roland, and Tabellini, 2007: 176)

The higher number of elected political parties in PR electoral systems leads to more coalition governments than is the case in plurality/majori-tarian electoral systems. Persson and his associates note that 63 percent of the election results in plurality/majoritarian electoral systems over this period were governed by a single party compared with only 17 percent for countries with PR electoral systems (Persson, Roland, and Tabellini, 2007).

4 Funk and Gathmann (2010) argued that it isn’t proportional representation itself, but “the political consequences, in particular the more fragmented legislature, that generates more spending in proportional systems.” This suggests that moving from a majoritarian electoral system that typically results in majority governments to a proportional system would likely lead to more fragmented legislatures, which in turn would lead to higher spending.

5 It’s worthwhile to note that a number of studies have concluded that the emergence of additional parties under PR electoral systems is a strategic response by politicians aimed at capturing additional political influence and power rather than a response to voter demands or strategic voting by the electorate. Persson, Roland, and Tabellini (2007) and Riker (1982) give additional information.



More recent data on elections in advanced industrial countries with different electoral systems provides further evidence to support the find-ings of Persson and his colleagues that plurality/majoritarian systems are correlated with lower levels of party fragmentation and lower incidences of coalition government. Table 1 summarizes the key results for a selection of industrialized countries, specifically, the average number of parties, the percentage of elections resulting in coalition governments, and the aver-age number of parties in coalitions based on election results between 2000 and 2015.

Figure 1 shows the average number of parties in the individual coun-tries6 included in this analysis. From 2000 to 2015, plurality/majoritarian electoral systems had an average number of effective parliamentary par-ties7 of 2.6 compared to 4.4 in PR systems8 (table 1 and figure 1). In other

6 The authors categorized the advanced industrialized countries used in this analysis based on information available from the Inter-Parliamentary Union (n.d.) and Norwegian Centre for Research Data (n.d.) databases.

7 The measure “effective parliamentary parties” is commonly used in comparative politics literature as a means for controlling for the effects of very small parties (for example a party that runs one candidate in one constituency) and parties that are unequal in size. The estimate is based on the seat share a party received in a given election. The estimator can be expressed as , where si is the seat share of the i-th party (see Lijphart, 1994).8 While electing some members to their legislatures through majoritarian voting, both Spain and Switzerland have been included as PR countries since only two of 350

N 1sS

i2�

�

Table 1: Electoral Outcomes by Electoral System, 2000-2015

Plurality/Majoritarian PR

Average Number of Effective Parties 2.6 4.4

Percentage of Coalition Elections 15% 83%

Average Number of Coalition Parties 3.0 3.3

Sources:: Norwegian Centre for Research Data (n.d.); Holger Döring and Philip Manow (2016); Inter-Parliamentary Union (n.d.); authors calculations.



Figure 1: Average Effective Number of Parliamentary Parties for Elections between 2000 and 2015

Sources: Norwegian Centre for Research Data (n.d.); Holger Döring and Philip Manow (2016);

Inter-Parliamentary Union (n.d.); authors calculations.

0 1 2 3 4 5 6 7 8

United States

United Kingdom

France

Avg. Plurality/Majoritarian

Australia

Canada

Turkey

Portugal

Spain

Greece

Ireland

Poland

Austria

Luxembourg

Italy Post-2005

Iceland

Czech Republic

Estonia

Avg. PR

Slovak Republic

Sweden

Norway

Slovenia

Switzerland

Finland

Denmark

Netherlands

Israel

Belgium

Plur

ality

/Maj

orita

rian

Prop

ortio

nal R

epre

sent

atio

n



words, countries with PR electoral rules had over 70 percent more effec-tive parliamentary parties, on average, than countries with plurality/ma-joritarian election rules. Although Canada has one of the highest number of political parties in the plurality/majoritarian system, it is likely that the number would increase even more with a PR system.

In terms of coalition governments, in advanced industrialized coun-tries with plurality/majoritarian electoral rules, only 15 percent of elec-tions between 2000 and 20159 resulted in coalition governments (table 1). This result was heavily influenced by Australia, in which three out of five elections in the period resulted in coalition governments. It is worth not-ing, however, that Australia does not have a first-past-the-post plurality electoral system, but rather an alternative vote majoritarian system.

In countries with PR election systems, on the other hand, 83 percent of elections resulted in coalition governments. Put differently, countries with PR election rules were 5.4 times more likely to have coalition governments than countries with plurality/majoritarian election rules.

Finally, the number of parties required to form coalition governments is higher in countries with PR election rules than in countries with plural-ity/majoritarian systems. On average, there were 3.4 parties in the coali-tions formed in countries with PR electoral rules compared to 3.0 par-ties in coalitions formed in countries with plurality/majoritarian electoral rules (table 1).

The idea that PR electoral systems elect more parties to the legislatures is not a controversial statement10 since PR is designed to allow more of the smaller, vote-receiving parties to be elected than plurality/majoritarian

members in the Spanish Congress of Deputies are chosen through majoritarian voting and only five out of 200 members in the Swiss National Council are elected through a majoritarian system. Elections on Italy were only included after 2005, when the country switched to a PR system.

9 Elections that resulted in another election due to the inability to form a government were not included in the analysis.

10 To read more about PR electoral rules leading to more parties, see Rae, 1971; Taagepera and Shugart, 1989; and Lijphart, 1990.



electoral systems.11 The link between these electoral rules and fiscal policy is referred to as a “common pool problem.” This key insight into the two electoral systems is often misunderstood and generally misrepresented in public debates.

Plurality/majoritarian systems provide the incentives for political par-ties to form coalitions within their party. In other words, the parties and their platforms must be broad enough to bring in a number of constitu-ency groups. An example of this is how major parties in most plurality/majoritarian countries handle voters who are focused on the environment. The incentive for each party is to have a strong enough platform to attract environmentally-motivated voters so there is no need for a stand-alone, environmentally-focused political party. In this way, the costs of the co-alitions are managed and incurred within the parties themselves. In the parlance of economics, the costs of the coalitions are largely internalized to the parties themselves.

PR systems, on the other hand, encourage the creation of stand-alone, single-issue, and regional parties. This means that instead of being part of a broader coalition within a single party, as is the case in plurality/majoritar-ian electoral systems, smaller, single-issue parties proliferate in countries with PR electoral systems. Those smaller parties contest elections along-side the larger, broader parties. Using the environmental example above, countries with PR systems tend to have single-issue environmental par-ties, most notably the Green Party, elected and represented in their Par-liaments.12

A side effect of the incentives embedded within PR electoral systems is that they are far more likely to require coalitions to govern since it is signif-icantly more difficult for them to secure a majority in an election. However, governing by coalition means that the costs of the coalition, which take the form of policies implemented in the government, are born not by the

11 For a more fulsome discussion of the links between electoral rules and party structures (as well as coalitions) see Persson, Roland, and Tabellini (2007).

12 While the Canadian legislature does contain a Green Party MP, the party has only ever elected a single MP in a legislature of over three hundred members.



party, as is the case with plurality/majoritarian systems, but by the broader public. Smaller, single-issue, and regional parties are able to exert propor-tionally much more power in government under PR electoral systems than under plurality/majoritarian electoral systems because the larger parties need their elected members to form a governing coalition. This propor-tionally greater power results in policies favoured by these smaller parties being enacted by government as a condition of the smaller parties’ support for the coalition in government. Indeed, one study of this dynamic showed that ministerial powers in coalition governments tended to be allocated to the parties with the strongest demonstrated preferences for those policy areas (Budge and Keman, 1990). The introduction of these policies by gov-ernment imposes costs on the electorate.

While electoral rules do not necessarily affect fiscal policy directly, the intermediate effect of more fragmented legislatures creates the incentives for greater public spending, which is financed in part by greater deficits.

III. Reviewing research on electoral systems and fiscal policy

This section summarizes the key research on the relationship between electoral rules (or electoral systems, as they are often called) and fiscal pol-icy, specifically the size of government spending and fiscal balance (deficits versus surpluses).

Size of government—Government spendingGiven the nature of the differences between plurality/majoritarian and PR electoral systems and their influence on the number of parties in Parlia-ment and the proclivity towards coalition governments, the key metric of how such differences affect fiscal policy is government spending.

In the countries discussed previously, there is a clear and observable relationship between electoral rules and government spending. Figure 2 uses IMF data to illustrate the average level of central government spend-ing from 2000 to 2014 in countries with plurality/majoritarian and PR



systems.13 Countries with PR electoral systems had average central gov-ernment spending of 29.2 percent of GDP compared to 23.5 percent for countries with plurality/majoritarian election rules. In other words, as a share of the economy (GDP), central governments in countries with PR election rules were almost one-quarter (24.3 percent) larger than in coun-tries with plurality/majoritarian election rules.

A number of academic studies have further developed the analysis of election rules and the size of government spending to account for other contributing factors that could affect government spending. Economists Yianos Kontopoulos and Roberto Perotti (1999) completed one of the first studies in this area. They examined the impact of government fragmenta-tion (i.e., more parties and more coalitions) on fiscal policy in 20 OECD countries between 1960 and 1995. They found that government spending increased as the number of coalition parties increased and that this impact was stronger in recessions.14 Milesi-Ferretti and her colleagues, publishing in the prestigious Quarterly Journal of Economics, similarly found that PR was associated with higher government spending in a sample of 20 OECD coun-tries15 between 1960 and 1995 (Milesi-Ferretti, Perotti, and Rostagno, 2002).

In their seminal 2003 book, The Economic Effect of Constitutions, Tor-sten Persson and Guido Tabellini examined the impact of electoral rules

13 Data on central government spending as a percentage of GDP come from the International Monetary Fund’s (2016) Government Finance Statistics (GFS) online data-base. The sector used is “central government excluding social security funds” which, for example, coincides with the National Accounts presentation for the federal government in Canada’s Fiscal Reference Tables. The analysis was carried out on the same advanced industrialized countries discussed above for which we examined differences in the aver-age number of parties and coalition governments. Countries were sorted into plural-ity/majoritarian and PR categories based on information from the Inter-Parliamentary Union (n.d.) and Norwegian Centre for Research Data (n.d.). Consistent with the above analysis, Spain and Switzerland were included as countries with PR systems.

14 Specifically, at zero GDP growth, Kontipoulos and Perotti (1999) found an extra party in a coalition added 0.12 percent of GDP per year to aggregate expenditure, but this increased by 0.02 percent of GDP for every one percentage point drop in GDP growth.

15 The study also examined 20 Latin American countries, but the evidence for a cor-relation between spending and electoral rules was weaker there than the relationship observed among OECD countries.



Figure 2: 2000-2014 Average, Central Government Expense as a percent of GDP

Sources: IMF (2016); Norwegian Centre for Research Data (n.d.); Inter-Parliamentary Union

(n.d.); authors calculations.

0 5 10 15 20 25 30 35 40

USA

Canada

France


Australia

UK

Switzerland

Spain

Slovakia

Poland

Turkey

Netherlands

Finland

Estonia

Italy

Luxembourg

Belgium

Czech Rep.

Avg. PR

Sweden

Slovenia

Iceland

Austria

Portugal

Norway

Ireland

Greece

Denmark

Israel

Plur

ality

/Maj

orita

rian

Prop

ortio

nal R

epre

sent

atio

n



on government spending in 85 countries. Their research demonstrated that central government spending was 5.7 percent of GDP lower in coun-tries with plurality/majoritarian compared to countries with elections held under proportional representation rules.16,17

Persson and Tabellini, along with their colleague Gerard Roland, fol-lowed up on their previous work with more sophisticated analytical tech-niques in a watershed essay that appeared in the Quarterly Journal of Political Science (Torsten, Roland, and Tabellini, 2007). The study’s conclu-sions buttressed their previous work; they determined that a switch from a plurality/majoritarian electoral system to a PR system would result in an increase in government spending of between 5 and 6 percent of GDP. Indeed, Persson, Roland, and Tabellini included a rather strongly worded conclusion regarding electoral systems and government spending:

… results strongly suggest that plurality/majoritarian elections indeed reduce public spending because they lead to a lower incidence of coalition governments. (Torsten, Roland, and Tabellini, 2007: 179)

16 Using another technique (two-stage least squares, or 2SLS), that did not yield signifi-cant results, the authors note that, “The fact that the estimated effects remain negative, large, and not too distant from the OLS estimates, reassures us of the validity of our inference, despite the large standard errors.”

17 A lengthy secondary literature has evaluated Persson and Tabellini’s work. Their find-ings on the impact of electoral rules have consistently held up to scrutiny, though there has been some debate over whether they have established causality (Acemoglu, 2005; Voigt, 2011). Their hypotheses have also been tested using alternative datasets, most notably Blume et al. (2009) who extended their model from 85 to 116 countries, also finding a significant correlation between majoritarian systems and lower government expenditures. Voigt (2011) was also able to replicate their results with respect to the correlation between majoritarian electoral systems and government spending. Rockey (2012) examined Persson and Tabellini’s hypotheses using both their original dataset and the extended dataset of Blume et al., also finding that majoritarian systems are associated with lower spending and fewer deficits (though noting that the results do not hold for countries that became democracies after the early 1990s). Similarly, Caruso et al. (2015) also found that the results hold for countries with a high degree of institutionalization (such as Canada), but not for less established democracies.



Finally, and more recently, Caruso and his colleagues examined Persson and Tabellini’s 2003 dataset and found that countries that held elections under plurality/majoritarian rules have central government expenditures that are 6.5 percent of GDP lower than countries that hold proportional representation elections.18

In addition to the impact on the size of government spending, different electoral rules also appear to have an effect on the composition of gov-ernment spending. Persson and Tabellini, for instance, found that moving from proportional to plurality/majoritarian electoral rules should lead to a 2 to 3 percent GDP decline in welfare spending (Persson and Tabellini, 2001).

The study led by Milesi-Ferretti referred to previously also found that governments elected under proportional systems tend to spend more on transfers, whereas governments elected under plurality/majoritarian sys-tems tend to purchase more goods and services (Milesi-Ferretti, Maria, Pe-rotti and Rostagno, 2002).

Fiscal balance—Deficits versus surplusesA second fiscal policy linked with electoral rules is how the higher levels of government spending in countries with PR electoral rules are financed. There are only two options to finance additional government spending: raise taxes and/or incur larger deficits, the latter of which simply defers the payment of taxes into the future. The preponderance of the evidence suggests that PR electoral systems lead to higher deficits.

Using the same methodology employed in the previous analysis of elec-tion rules and government spending, figure 3 illustrates the average deficit

18 Caruso, Scartascini, and Tommasi (2015) defined institutionalization as the degree to which political power is exercised through “formal political arenas such as the legislature or the political party system.” Focusing on countries with high institutionalization, such as Canada, rather than on countries with low institutionalization is important, since the nuances of political arrangements are less salient in countries where political power can be exercised outside of formal legislative channels (i.e., where democratic institutions aren’t always respected). Indeed, the authors found that using Persson and Tabellini’s database, they were able to replicate the results for countries with high institutionaliza-tion, but the results were not significant for those with low institutionalization.



or surplus for the central governments of various countries with either plurality/majoritarian or PR election rules between 2000 and 2014. The overall finding is that countries with PR election rules actually maintained lower average deficits (-1.5 percent of GDP) than countries with plurality/majoritarian election rules (-2.0 percent of GDP). However, these results are inordinately influenced by Norway’s large surpluses (14.2 percent of GDP) over this period (Figure 3), which are largely a function of its in-dependent sovereign fund.19 If Norway is removed from the analysis, the average deficit for PR countries increases to 2.2 percent of GDP, roughly 10 percent higher than the average deficits for countries with plurality/ma-joritarian election rules.

A great deal of academic work has extended the analysis of election rules and fiscal balance to include other pertinent variables. Grilli, Masci-andaro, and Tabellini, for instance, examined fiscal outcomes in 18 OECD countries between 1950 and 1989. They found that all but one of the coun-tries with representational electoral systems20 had “unsustainable fiscal policies” (Grilli, Masciandaro, Tabellini, Malinvaud, and Pagano, 1991). They concluded that “representational democracies” were more likely to have public debt problems than countries relying on plurality/majoritarian electoral rules.

In their 2003 book on this subject, Persson and Tabellini examined 85 democracies between 1960 and 1998 and found that plurality/majoritar-ian elections resulted in roughly 2 percent lower deficits than elections under proportional representation. In a subsequent paper that extended their work further, they concluded that increased spending under propor-tional electoral systems is “financed by higher taxes and deficits in similar proportions” (Persson and Tabellini, 2004a).

As noted in the conceptual section of this chapter, coalition govern-ments are much more likely to be formed under PR election rules than

19 For information on Norway’s sovereign wealth fund, see Murphy and Clemens (2013).

20 The authors grouped countries with five or fewer representatives per district as majoritarian, and those with more than five as representational. While the definition of majoritarian here is broader than the typical definition, it highlights that proportionality erodes the incentives for fiscal discipline.



Figure 3: 2000-2014 Average, Central Government Net Operating Balance as a percent of GDP

Sources: IMF (2016); Norwegian Centre for Research Data (n.d.); Inter-Parliamentary Union

(n.d.); authors calculations.

-10.0 -5.0 0.0 5.0 10.0 15.0

UK

France

USA


Australia

Canada

Greece

Portugal

Israel

Slovakia

Ireland

Poland

Italy

Slovenia

Czech Rep.

Avg. PR (excluded Norway)

Spain

Austria

Belgium

Iceland

Finland

Turkey

Netherlands

Switzerland

Luxembourg

Denmark

Sweden

Estonia

Norway

Plur

ality

/Maj

orita

rian

Prop

ortio

nal R

epre

sent

atio

n



under plurality/majoritarian rules. Analyzes of the link between the type of election rules and the presence and severity of deficits has generally linked PR election rules to more frequent and larger deficits. For instance, noted economists Nouriel Roubini and Jeffrey Sachs analyzed the impact of minority governments in 15 industrialized democracies between 1961 and 1985. They concluded that minority governments have 1.2 percent-age points of additional budget deficits relative to majority governments (Roubini and Sachs, 1989).

Alesina and Perotti provided further evidence for this concept in a 1995 study. They examined fiscal adjustments in 20 OECD governments be-tween 1960 and 1992 and found that “conflicts amongst coalition members and the fragility of coalition governments make it difficult to maintain a ‘tough’ fiscal stance, particularly when politically sensitive programs, gov-ernment employment and social security are involved.” They concluded that this rendered coalition governments “generally unable to carry out successful fiscal adjustments” (Alesina and Perotti, 1995: 24).

Similarly, scholars Kontopoulos and Perotti, referred to earlier, exam-ined the impact of government fragmentation on budget deficits in 20 OECD countries from 1960 to 1995. They found a significant positive cor-relation between the number of parties within a governing coalition and public debt between 1984 and 1995 (Kontopoulos and Perotti, 1999).21

They also found that in “bad times,” larger coalitions have more power-ful negative effects. Specifically, they found that elections under propor-tional representation tended to require coalition governments more fre-

21 The authors pointed out that the 1960s were characterized by “more or less stable growth and little fiscal action in most countries of the sample,” which likely explains why there was no correlation between electoral rules and deficits during that period. They also pointed out that the fiscal challenges during the 1970s were due to external shocks common to all countries, which meant that the “executive decision-making process” was most important for constraining spending growth. In the 1980s “the dividing line was between those countries that engaged in large discretionary consolidations and those that did not.” They argued that the number of coalition partners was a greater factor in the 1980s because “fiscal consolidation is largely political and requires a cohesive government” (Kontopoulos and Perotti, 1999: 91).



quently than plurality/majoritarian systems, suggesting that moving to PR would lead to higher government expenditures and worse deficits during negative economic shocks.

Finally, one of many studies on this issue by scholars Persson and Ta-bellini also found evidence that the challenges governments elected under proportional systems have in balancing their budgets may linger beyond economic downturns. Specifically, they found evidence that the type of electoral rule is correlated with the reaction of government to economic shocks. In countries with PR electoral rules, spending as a share of GDP rises in recessions but does not decline in booms, while cyclical fluctua-tions tend to have symmetric impacts on fiscal policy under other electoral systems (Persson and Tabellini, 2004a: 24–45).

The totality of the evidence strongly suggests that governments elected by proportional representation are more likely to run deficits than those elected through plurality/majoritarian systems. This is largely due to the increased likelihood of electing more parties that necessitate coalition governments.

IV. Conclusion

There has been a great deal of research into the impact of electoral rules on fiscal policy over the last several decades. Many of these studies have focused on the relationship between electoral systems and public spending and deficits.

The research shows that proportional electoral rules are linked with higher public spending than plurality/majoritarian systems. Moreover, the literature demonstrates a clear connection between proportional repre-sentation and deficits. Finally, the literature suggests that more fractured a government is (with larger, unstable coalitions), the more difficult a time it will have responding to fiscal crises.

The tendency of proportional electoral systems to elect minority govern-ments is a serious consideration when weighing the benefits and costs of various electoral systems. Plurality/majoritarian electoral systems, by con-



trast, typically elect majority governments. The literature clearly suggests that a move from our plurality/majoritarian electoral system to a propor-tional system would likely increase both government spending and deficits.

References

Acemoglu, Daron (2005). Constitutions, Politics, and Economics: A Review Essay on Persson and Tabellini’s The Economic Effects of Constitutions. Journal of Economic Literature 43, 4 (December): 1025–1048.

Alesina, Alberto, and Roberto Perotti (1995). Fiscal Expansions and Fiscal Adjustments in OECD Countries. NBER Working Paper No.5214 (August). National Bureau of Economic Research.

Blume, Lorenz, Jens Muller, Stefan Voigt, and Carsten Wolf (2009). The Economic Effects of Constitutions: Replicating—and Extending—Persson and Tabellini. Public Choice 139, 1-2 (April): 197–225.

Caruso, German, Carlos Scartascini, and Mariano Tommasi (2015). Are We All Playing the Same Game? The Economic Effects of Constitutions Depend on the Degree of Institutionalization. European Journal of Political Economy 38 (June): 212–228.

Döring, Holger, and Philip Manow (2016). Information on parties, elections, and cabinets in modern democracies. Parliaments and Governments Database. Development version. ParlGov. <http://www.parlgov.org/> , as of June 22, 2016.

Funk, Patricia, and Christina Gathmann (2010). How Do Electoral Systems Affect Fiscal Policy? Evidence from State and Local Governments, 1890 to 2005. CESifo Working Paper Series No. 2958. Center for Economic Studies. <https://www.aeaweb.org/conference/2009/retrieve.php?pdfid=131>, as of June 22, 2016.

http://www.parlgov.org/

https://www.aeaweb.org/conference/2009/retrieve.php?pdfid=131



Grilli, Vittorio, Donato Masciandaro, Guido Tabellini, Edmond Malinvaud, and Marco Pagano (1991). Political and Monetary Institutions and Public Financial Policies in the Industrial Countries. Economic Policy 6, 13 (October): 341–92.

International Monetary Fund [IMF] (2016). Government Finance Statistics. Web portal. IMF. <http://data.imf.org/?sk=a0867067-d23c-4ebc-ad23-d3b015045405&sId=1390030109571>, as of June 22, 2016.

Inter-Parliamentary Union (n.d.). PARLINE Database on National Parliaments. Web database. Inter-Parliamentary Union. <http://www.ipu.org/parline-e/parlinesearch.asp>, as of June 22, 2016.

Kontopoulos, Yianos, and Roberto Perotti (1999). Government Fragmentation and Fiscal Policy Outcomes: Evidence from the OECD countries. In J. Poterba and J. von Hagen (eds.). Fiscal Institutions and Fiscal Preference (University of Chicago Press): 81–102.

Liberal Party of Canada (2015). Real Change: A New Plan for Strong Middle Class. Liberal Party of Canada. <https://www.liberal.ca/files/2015/10/New-plan-for-a-strong-middle-class.pdf>, as of June 22, 2016>.

Lijphart, Arend (1990). The Political Consequences of Electoral Laws, 1945–85. American Political Science Review 84, 2 (June): 481–496.

Lijphart, Arend (1994). Electoral Systems and Party Systems: A Study of Twenty-Seven Democracies, 1945-1990. Oxford University Press.

Lizzeri, Alessandro, and Nicola Persico (2001). The Provision of Public Goods under Alternative Electoral Incentives. American Economic Review 91: 225–245.

MacRae, C. Duncan (1977). A Political Model of the Business Cycle. Journal of Political Economy 85, 2 (April): 239–263.

http://data.imf.org/?sk=a0867067-d23c-4ebc-ad23-d3b015045405&sId=1390030109571

http://data.imf.org/?sk=a0867067-d23c-4ebc-ad23-d3b015045405&sId=1390030109571

http://www.ipu.org/parline-e/parlinesearch.asp

http://www.ipu.org/parline-e/parlinesearch.asp

https://www.liberal.ca/files/2015/10/New-plan-for-a-strong-middle-class.pdf

https://www.liberal.ca/files/2015/10/New-plan-for-a-strong-middle-class.pdf



Milesi-Ferretti, Gian Maria, Roberto Perotti, and Massimo Rostagno (2002). Electoral Systems and Public Spending. Quarterly Journal of Economics 117, 2: 609–657.

Murphy, Robert P., and Jason Clemens (2013). Reforming Alberta’s Heritage Fund: Lessons from Alaska and Norway. The Fraser Institute. <https://www.fraserinstitute.org/sites/default/files/reforming-albertas-heritage-fund.pdf>, as of June 22, 2016.

Norwegian Centre for Research Data (n.d.). European Elections Database. Norwegian Centre for Research Data. <http://www.nsd.uib.no/european_election_database/about/>, as of June 10, 2016.

O’Neal, Brian (1993). Electoral Systems. Canada, Library of Parliament. <http://www.lop.parl.gc.ca/content/lop/researchpublications/bp334-e.htm>, as June 22, 2016.

Persson, Torsten, and Guido Tabellini (2001). Political Institutions and Policy Outcomes: What are the Stylized Facts? CEPR Discussion Paper No. 2872. Centre for Economic Policy Research.

Persson, Torsten, and Guido Tabellini (2003). The Economic Effects of Constitutions: What Do the Data Say? MIT Press.

Persson, Torsten, and Guido Tabellini (2004a). Constitutional Rules and Fiscal Policy Outcomes. American Economic Review 94, 1 (March): 24–45.

Persson, Torsten, and Guido Tabellini (2004b). Constitutions and Economic Policy. Journal of Economic Perspectives 18, 1: 75-98.

Persson, Torsten, and Guido Tabellini (2008). Electoral Systems and Economic Policy. In Donald A. Wittman and Barry R. Weingast (eds.). The Oxford Handbook of Political Economy (June).

https://www.fraserinstitute.org/sites/default/files/reforming-albertas-heritage-fund.pdf

https://www.fraserinstitute.org/sites/default/files/reforming-albertas-heritage-fund.pdf

http://www.nsd.uib.no/european_election_database/about/

http://www.nsd.uib.no/european_election_database/about/

http://www.lop.parl.gc.ca/content/lop/researchpublications/bp334-e.htm



Persson, Torsten, Gerard Roland, and Guido Tabellini (2007). Electoral Rules and Government Spending in Parliamentary Democracies. Quarterly Journal of Political Science 2, 2: 155–188.

Rae, Douglas (1971). The Political Consequences of Electoral Laws, 2nd ed. Yale University Press.

Reynolds, Andrew, Ben Reilly, and Andrew Ellis. (2008). Electoral System Design: The New International IDEA Handbook. International Institute for Democracy and Electoral Assistance. <http://www.idea.int/publications/esd/>, as of June 22, 2016.

Riker, William (1982). The Two-Party System and Duverger’s Law: An Essay on the History of Political Science. American Political Science Review 76, 4 (December): 753–766.

Rockey, James (2012). Reconsidering the Fiscal Effects of Constitutions. European Journal of Political Economy 28, 3: 313–323.

Roubini, Nouriel, and Jeffrey D. Sachs (1989). Political and Economic Determinants of Budget Deficits in the Industrial Democracies. European Economic Review 33, 5: 903–933.

Taagepera, Rein, and Matthew Shugart (1989). Seats and Votes: The Effects and Determinants of Electoral Systems. Yale University Press.

Voigt, Stefan (2011). Empirical Constitutional Economics: Onward and Upward? Journal of Economic Behaviour & Organization 80, 2 (October): 319–330.

http://www.idea.int/publications/esd/



About the authors

Jason Clemens is the Executive Vice President of the Fraser Institute. He has

an Honors Bachelors Degree of Commerce and a Master’s Degree in Business

Administration from the University of Windsor as well as a Post Baccalaureate

Degree in Economics from Simon Fraser University. He has published over 70 major

studies on a wide range of topics, including taxation and entrepreneurship. He has

published over 300 shorter articles in US, Canadian, and international newspapers.

Taylor Jackson is a Senior Policy Analyst at the Fraser Institute. He holds a B.A.

and an M.A. in Political Science from Simon Fraser University. Mr. Jackson is the

co-author of a number of studies and book chapters on a wide range of policy issues,

including energy, mining and natural resources, democratic reform, Canadian-

American relations, and other areas. Mr. Jackson’s work has been covered in the

media around the world and his commentaries have appeared in the National Post,

Financial Post, and Washington Times, as well as other newspapers across Canada.

Steve Lafleur is Senior Policy Analyst at the Fraser Institute. He holds an MA

in Political Science from Wilfrid Laurier University and a BA from Laurentian

University where he studied Political Science and Economics. His past work has

focused primarily on housing, transportation, local government, and inter-govern-

mental fiscal relations. His current focus is on economic com petitiveness of juris-

dictions in the Prairie provinces.

Joel Emes is Fraser Institute Senior Fellow who rejoined the Institute after a stint as

a senior advisor to British Columbia’s provincial government. He previously served

as a senior analyst, then as executive director (2009 to 2011), at the BC Progress

Board. Prior to that, Joel was a senior research economist at the Fraser Institute,

where he initiated and led several flagship projects in the areas of tax freedom and

government performance, spending, debt, and unfunded liabilities. Joel holds a BA

and an MA in economics from Simon Fraser University.



AcknowledgmentsThe authors would like to express their appreciation to the Lotte and John Hecht

Memorial Foundation for their support of this project. They would also like to rec-

ognize the research assistance of Kayla Ishkanian and Matthew Lau. They also thank

the unidentified reviewers for their many helpful comments and suggestions. Any

remaining errors or oversights are the sole responsibility of the authors. As the

researchers have worked independently, the views and conclusions expressed in

this paper do not necessarily reflect those of the Board of Directors of the Fraser

Institute, the staff, or supporters.


CHAPTER 3 The Consequences of the Alternative VoteLydia Miljan and Taylor Jackson1

[I]nstitutions matter—in this case, the institutions by which we elect people to office. Institutions affect outcomes by framing the choices we make. If we care about the policy outcomes that are the results of those choices, it is nothing less than folly to ignore the institutions through which policies are developed and decided.

—Henry Milner, 1999: 18

In its frequently asked questions web page about electoral reform, the federal government categorically states, “In a multi-party democracy like Canada’s, FPTP [first-past-the-post] distorts the will of the elector-ate” (Canada, n.d.). The web page also claims that this distortion explains low voter turnout and engagement. Although the government admits that changes in electoral reform are far reaching, apart from its commitment to change the system, the government itself has declared that it has no

1 Lydia Miljan is a senior fellow and Taylor Jackson is an analyst with the Fraser Institute. Please see the end of this chapter for their full biographies. The authors are grateful to Joel Emes for his assistance in developing the alterna-tive vote simulation’s methodology. The authors also wish to thank Kyle Sholes, Kayla Ishkanian, and Matthew Lau for their research assistance.



preference and that it wants an all-party committee it has struck to de-liver a recommendation as to how the system should be changed. On that committee are members of the Green and NDP parties, whose campaign platforms in the 2015 election campaign called for proportional represen-tation. Much like the Liberals, the Conservatives have not stated a pref-erence for one system over another. Their position is that whatever the committee decides, the final say should rest with public in the form of a referendum. Despite Prime Minister Trudeau’s reassurance that he will not try to influence the outcome of the deliberations, he has previously indi-cated that his preferred model is the ranked ballot method (sometimes re-ferred to as alternative vote, or AV). Political parties use a similar method to elect their leaders.

The simplicity of the AV is one element that could explain the prime minister’s attraction to this system. Unlike a mixed member system, or proportional representation, which would require the electoral boundaries to be redrawn, or seats added to the House of Commons, AV entails only minor revisions to our electoral system. Essentially all that would change would be the ballot itself, which would change from one vote marked by an X, to a rank ordering of the ballot. An additional complexity added to the system is the counting of the votes in ridings where no candidate won 50-percent-plus-one of the votes.

The government has given the committee five “guiding principles” to help their deliberations:

1 Restore the effectiveness and legitimacy of voting by reducing distortions and strengthening the link between voter intention and the electoral result.

2 Encourage greater engagement and participation in the demo-cratic process, including inclusion of underrepresented groups.

3 Support accessibility and inclusiveness for all eligible voters while avoiding undue complexity in the voting process.

4 Safeguard the integrity of our voting process.5 Preserve the accountability of local representation.


The Consequences of the Alternative Vote d 47

It is not clear how the AV system would fulfill any of these principles more effectively than our current first-past-the-post (FPTP) system. In fact, one of the reasons for the simplicity of the AV system is that it is within the same electoral family as the FPTP system. Both are considered majority/plurality systems (see table 1). As John Pepall has discussed else-where in this volume, AV systems create a majority by allowing people who chose as their first choice a party or candidate that did not win suf-ficient votes to stay in the race, the ability to exercise their vote a second time, by having their second or subsequent choice be counted. This creates a majority, despite the fact that the winning party did not have the most first-choice selections. Some have suggested that this manufactures a ma-jority where one did not exist in the absence of the ranked ballot (Fair Vote Canada, n.d,). Moreover, AV does not provide proportionality (Lijphart, 1994; Flanagan, 1999; Jansen, 2004). However, what AV does is eliminate the so-called “wasted” votes and could reduce strategic voting (Kay, 2016).2

2 A “wasted vote” typically refers to votes that are cast for a candidate who has a very small chance of winning. It can also refer to votes over and above that which a candi-date needed to receive a seat. “Strategic voting” occurs when a voter casts a vote for a

Table 1: Comparision of Electoral Systems

Category Type of System Countries (examples)

Majority/ Plurality

Single-Member Plurality (SMP), often termed First-Past-The-Post (FPTP)

Canada, India, UK, USA

Alternative Vote (AV) Australia, Papua New Guinea

Two-Round System (TRS) Cuba, France, Haiti, Vietnam

Proportional Representation

List-PR Belgium, Greece, Israel, Spain, Portugal

Single Transferable Vote (STV) Ireland, Malta

Mixed Mixed Member Proportional (MMP) Germany, Mexico, New Zealand, Romania

Sources: Institute for Electoral Democracy and Electoral Assistance (2010). Table of Electoral Systems Worldwide. <http://www.idea.int/esd/world.cfm>, as of August 29, 2016.



Given the prime minister’s preference, there have been some attempts to speculate about how an AV system would have affected the outcome of the last election (Grenier, 2015a, Jansen, 2015). As Ken Carty aptly notes, it is difficult to predict how a change to the electoral system will affect future elections:

Rules that change the voting system will directly affect the whole sys-tem, including the way candidates are chosen and who chooses them, the organizing and financing of campaigns and how money is spent and collected in campaigns, the structure and the internal dynamic of political parties, the number and the character of the political parties in the electoral contest, the number and the character of the political parties that get elected to the House of Commons, and, of course, the decision-making mechanisms that voters go through when they choose. (Carty, 2016)

Although we cannot know how elections in the past may have been-affected by these strategic decisions, we can, nonetheless, see whether a given system, in theory, has an undue effect on the political system. Ex-ercises such as those performed by Grenier and Jansen tell us something about which party benefits the most as well as which parties would have to change their strategies to accommodate the new system. This paper ex-pands that analysis to examine not just the last federal election, but all federal elections from 1997 to 2015.

Overview of the Alternative Vote system

AV is a majoritarian voting system that requires members elected to the legislature to have received a majority of the votes cast in a given constitu-

candidate based on their evaluations of potential electoral outcomes. For example, in a constituency with three candidates—A, B, and C—a voter whose first preference is C may instead vote for B if C has a minimal chance of getting elected.



ency. The system is similar to first-past-the-post in that voters elect one member per constituency.

AV differs from FPTP in that when voters go to the polls, they rank the candidates running in their constituency based on their preferences. When voting concludes, the first preferences are tallied up and any candi-date receiving a majority of first preference votes is elected to the seat in their constituency. If no candidate receives a majority of the first prefer-ence votes, then the candidate with the fewest votes is dropped and their votes are redistributed based on the second or subsequent preferences of their voters. If there is still no majority after the first round of elimination, the process is repeated until one candidate does achieve a majority.

This voting system has been most widely used in Australia, where it was adopted in 1918 and has been the system used to elect members to the lower house in the Australian national legislature since its first use in the election of 1919.3 The system was introduced by a non-Labour4 gov-ernment in response to the formation of the Country Party, a competing right-of-centre party that was establishing a growing regional influence. Implementing the AV electoral system allowed the non-Labour parties to exchange voting preferences, particularly in rural districts. This helped to prevent the Labour Party from benefitting when the non-Labour vote was split (Sharman et al., 2002).

The Country Party has benefited the most from the implementation of the AV electoral rules, as the party has often received a seat share greater than its vote share. Indeed, non-Labour parties have tended to be the ben-eficiaries of this system. In her examination of the effects that AV had on Australian elections from 1919 to 1951, Rydon (1956) found that of the 73 seats that required a distribution of preferences to determine the winner, 58 of them were won by non-Labour candidates and only 15 were won by Labour candidates, demonstrating the possible impact that the implemen-tation of this system had on Labour’s electoral prospects.

3 In Australia, it is mandatory that voters both vote and indicate a preference for all candidates.4 The Labour Party is the dominant left-of-centre party in Australia.



The alliance between the successor to the Country Party, the Nation-al Party of Australia, and the dominant right-of-centre party, the Liberal Party, still exists, leading to the creation of coalition governments in cases where the two parties have a seat share greater than that of the Labour Party. While the early beneficiaries of AV were the right-of-centre parties, more recently, as is discussed below, the Labour Party has been able to capitalize on some of the dynamics that AV creates, such as the exchange of voting preferences, although this has not resulted in the existence of formal coalition governments led by the Labour Party.

Bean (1997) has described some of the other effects that AV has had on Australia’s political dynamics. One of Bean’s (1997) important insights is that AV makes minor parties in general more influential because they can play a crucial role in determining which of the major parties receives the seat in close constituencies by shaping how individuals who have a first preference for the minor party order their subsequent preferences. Repre-sentatives from minor parties often distribute “how to vote” cards to their supporters, indicating how to structure their preferences.

In addition to the instances discussed above, Bean (1997) recounts a number of instances in Australian political history where the influence of minor parties has shaped electoral outcomes. In the 1950s and ’60s, the Democratic Labour Party was the most popular minor party in the country and their supporters’ subsequent preferences tended to heavily favour the Liberal and National parties.5 By the 1980s the Australian Democrats had become the dominant minor party in the country and their supporters’ subsequent preferences tended to benefit the Labour party.

Regarding the role of minor parties, Bean (1997) also describes an in-teresting dynamic that took place during the 1990 election. During the campaign period, the Labour party adopted a strategy of trying to maxi-mize their second choice vote share, after acknowledging that many voters would probably prefer the Australian Democrats or other minor parties. Bean notes that the Labour Party’s adoption of the “second preference strategy” was credited with winning the Labour Party the 1990 election.

5 The Liberal and National parties have a long-standing coalition.



The influence that minor parties can have under AV is in some ways similar to the influence that minor parties have in proportional represen-tation (PR) systems (Clemens et al., 2016). In both systems, larger parties need the support of minor parties to control government. In AV, the larger parties need the second and subsequent preference votes from support-ers of minor parties in order to win tight races, while in PR systems, large parties typically need the support of minor party representatives in order to form coalition governments. In both situations, it is possible that these dynamics allow minor parties to assert parts of their agendas, which leads to them having a disproportionate amount of influence. However, the in-fluence of minor parties is likely greater in PR systems where the support of minor parties is consistently needed in order for large parties to govern.

Another issue with AV is that it can lead to higher levels of invalid votes than FPTP. Two of the reasons for this are 1) that voting in the AV system is more complicated for the voter, and 2) some voters may not wish to ex-press a preference for certain parties even though they are required to do so, as is the case in Australian federal elections. The higher level of rejected or invalid ballots can be seen by comparing a recent Australia and Cana-dian election. In the 2015 Canadian election, only 0.7 percent of ballots were rejected (Elections Canada, 2015). In the 2013 Australian election, 5.9 percent of votes were considered to be invalid in House of Representa-tives voting (Australian Electoral Commission, 2015).

So called “donkey voting” is another, smaller concern that results from AV voting in Australia. It occurs when voters simply write successive num-bers starting at the top of the ballot without any concern for who the par-ticular candidates are (Bean, 1997). Only a very small percentage of voters do this, and the proportion who do so would not differ much from those individuals who select a candidate at random in a FPTP election.

Canada’s history with the AV

It has been well established that no electoral system is ideal, as Australia’s experience with AV shows, and that changes to the system will invariably



help some parties at the expense of others. However, designing an elec-toral system to ensure that one party remains in power (or that another party is kept out of power), is not only unfair, but potentially dangerous. Western Canadian provinces have experimented with the AV system in part to achieve some of these nefarious goals. For example, in British Co-lumbia, AV was adopted by the coalition of Liberals and Conservatives to prevent the Cooperative Commonwealth Federation (CCF) from gaining power in 1952. The Liberals and Conservatives had hoped that each would be the other’s second choice, and therefore continue the coalition between them without having to go through the trouble of merging the two par-ties. Ironically, the attempt backfired when Social Credit won a minority government in the 1952 election because the vote transfers benefitted So-cial Credit over the other parties, especially the CCF (Elkins, 1976). This repeated itself in the 1953 election where Social Credit won a majority government and then abandoned AV entirely.

Harold Jansen (2004) uses the British Columbia elections as well as those conducted in Alberta and Manitoba under AV to examine the real world impact on electoral issues of proportionality, number of political parties, voter turnout, and spoiled ballots. He found that the AV system had no effect on proportionality or the number of political parties, or even voter turnout. However, it did produce a higher number of spoiled or re-jected ballots, something to be expected given Australia’s experience, es-pecially in jurisdictions that required that all ballots be completed with all preferences listed in rank order, and not just an X marked by an indi-vidual’s first choice. Jansen’s work underscores the fact that each election is unique as is the political environment in which elections are fought. With the caveat that had prior elections been held under different rules, political parties would have behaved differently, we seek to find out whether there is a systematic bias in favour of one party over the others given the regional and party history in which Canadian elections have been fought during the last two decades.



Alternative Vote Simulation—Data and Methodology

DataTo conduct the simulation, we collected data for the 1997 to 2015 period on Canadian federal election outcomes and voter preferences that went beyond those voters’ first choice. These data come from the Constituen-cy-Level Elections Archive (CLEA) (Kollman et al., 2016). The CLEA is a repository that collects data on lower-house elections for 136 countries and territories. We transformed the data by removing any candidates and parties that had less than five percent of the vote in a given constituency and/or which did not have subsequent choice data. This means that some Green Party candidates who received less than five percent of the vote in their constituencies were still included in the analysis in years where we had data on the second choice preferences of Green Party voters. Most of the parties eliminated in this process had vote shares of one percent or less, and, as an example, in the 2015 election this affected only 0.6 percent of votes.

The data on subsequent choice preferences come from the Canadian Election Studies (CES) (Blais et al., 1997, 2000, 2004, 2006; Gidengil et al., 2008; Fournier et al., 2011, 2016). The CES provides data on second-choice vote preferences that have been collected with a consistent methodology over time. Recent iterations of the CES are composed of four surveys: a campaign period survey, a post-election survey, and separate subsequent mail back and web-based surveys. Our analysis used the results of the cam-paign period survey only, so as not to run the risk of having second-prefer-ence data biased by knowledge of the outcome of the election. The survey’s sample is designed to be representative of the Canadian adult population; it is composed of Canadian citizens over the age of 18 who speak either English or French and reside in a private residence in the 10 Canadian provinces (territories are excluded).6 The lowest level of disaggregation possible is at the provincial level, so this is the level used throughout the

6 For more information on the methodology of the CES, see Northrup (2012).



simulation. Any “don’t know” or “no second choice” responses were re-moved. As an example, of the voters who indicated a first preference for a party, about 28 percent of them indicated that they “don’t know” their second choice or simply had “no second choice.” These types of responses were more likely to come from respondents who indicated that their first choice was for the Conservatives than other parties.

MethodologyWe estimated the results for the seven Canadian federal elections that oc-curred between 1997 and 2000, following an alternative vote (AV) elec-toral system, using an iterative approach. Constituencies where the lead candidate had at least 50 percent of the vote were not changed. In all other constituencies, the votes cast for the last place party were redistributed based on the second-choice data described above. If this initial redistribu-tion did not produce a clear winner, the process was repeated. The process continued until a candidate received at least 50 percent of the vote.

Due to the unavailability of a consistent data series that has voting pref-erences beyond a second choice, we assume that when a party is elimi-nated their votes are redistributed following the second choice preferences of individuals who indicated a first choice preference for that party, even if votes had already been redistributed to that party.

There were also a number of special cases resulting from missing CES second-choice data. These special cases were estimated in the following manner:

• No data were available for the territories. Election results were re-estimated in the territories by using the national averages.

• In cases where independents or other parties had a vote share of five percent or more and were therefore included in the AV re-estimations, their votes were redistributed based on the aggregate provincial second-choice averages. To distribute votes to indepen-dent or other parties that had not been eliminated, we distributed votes from the eliminated party equal to the initial vote share of the independent or other party.



• Second-choice data for the Green Party was sparse in the 1997 and 2000 election. In addition, the average vote shares for the Green Party were quite small at 1.5 percent in 1997 and 2.1 percent in 2000. This compares to around four percent in the subsequent elections. For these reasons, any Green Party candidates with less than five percent of the vote were removed from the analy-sis. In the few BC cases in those two elections where Green Party candidates had an initial vote share of more than five percent, upon elimination their votes were redistributed via the aggregate second-choice averages for their province.

• In cases where there were no provincial data for Green Party or Reform Party second choices, national second-choice data were used to redistribute votes to and from candidates representing these parties.

• In 1997, there were no data for the Reform Party available in Quebec. In order to redistribute their votes, Canadian Alliance data from Quebec from the 2000 election were used.

• In 2004, there were no data available based on a first-choice preference of the Green Party. In order to redistribute Green Party votes, 2006 data were used.

Results

This section presents the results of the AV simulation of the seven previous federal elections. After presenting some of the general statistics on how an AV election might compare to the current FPTP system, we will pres-ent the results of each re-estimated election. The appendix gives detailed results by province for each of the seven elections.

Table 2 presents some of the general outcomes of the analysis. On aver-age, for the seven federal elections, in 56 percent of the constituencies a candidate did not receive a majority of votes, meaning that the AV process had to be implemented. While this may seem high, the number of constit-uencies where the AV electoral system would have resulted in a different



party’s candidate being elected was much smaller. On average, 11 percent of seats went to a different candidate when AV was used. This ranged from a high of 17 percent of seats going to different candidates in 2015, to six percent of seats being changed in 2000. Our estimates also indicate that in three of the seven elections, an AV electoral system could have resulted in a different party winning the election, or could have changed whether the winning party had a majority or minority government.

In the 1997 election, our estimates indicate that the result would have been the same under either the FPTP or AV system—a Liberal majority (table 3). Under an AV system, the Liberals and Progressive Conservatives (PCs) were the only two parties that would have gained additional seats. The PCs would have added six seats: two in Quebec from the Bloc Quebe-cois (BQ), one in New Brunswick and Newfoundland each from the Liber-als, and two in Manitoba from the Reform Party. The Liberals’ gain of 15 seats under the AV system was primarily centered in Quebec, where the party gained seven seats, all from the BQ, and in British Columbia where they also gained seven seats, most of which came from the Reform Party.

Table 2: Seats in Which Redistributed and Results Changed, 1997-2015

Seats Requiring Redistribution Seats with Changed Results

Number Percent Number Percent

1997 185 61% 26 9%

2000 129 43% 19 6%

2004 168 55% 40 13%

2006 182 59% 41 13%

2008 185 60% 26 8%

2011 158 51% 37 12%

2015 201 59% 58 17%

Note: During the 1997 and 2000 elections there were 301 seats in the House of Commons. This

was increased to 308 seats from 2004 to 2011. In 2015 election, 338 seats were contested.



While the PCs and Liberals would have gained a number of seats, the ma-jor positions in Parliament did not change. The Reform Party remained as the Official Opposition after estimating the outcome of the 1997 election using AV rules, and the BQ remained as the party with the third highest number of seats. The only change in the position of the parties was that the PCs moved into being the party with the fourth highest number of seats, displacing the NDP.

Calculating seat totals for the 2000 election using the AV system results in similar outcomes to the previous election (table 4). The results of the election remained the same with the Liberals receiving a majority govern-ment. The majority was strengthened under AV with the Liberals receiving 62 percent of the seats in the House of Commons using these voting rules compared to 57 percent of seats under FPTP. Under AV, the Liberals also saw the largest gains in seat totals, increasing their take to 187 seats, up from 172 seats. The only other party to gain seats was the BQ who saw their seat total increase by one. The gains for the Liberals were distributed throughout many different provinces and all parties. As with the previous election, the Canadian Alliance would have remained as the Official Op-position, and the only change in the power of parties was between the PCs and NDP.

Table 3: 1997 Results Comparison

Reform PC Green Liberal NDP Bloc Indepen- dent

FPTP AV FPTP AV FPTP AV FPTP AV FPTP AV FPTP AV FPTP AV

Total Seats

Won

60 49 20 26 0 0 155 170 21 20 44 35 1 1

% of Seats

Won

20% 16% 7% 9% 0% 0% 51% 56% 7% 7% 15% 12% 0% 0%

Result FPTP: Liberal Majority

Result AV: Liberal Majority



The 2004 election is the first in our sample where we see a change in the outcome (table 5). Had this election been conducted under AV rules, the Liberals would likely have received a majority instead of a minority govern-ment. Specifically, the Liberals would have gone from having 44 percent of seats (135 seats) in the House of Commons under FPTP to 52 percent of the seats (160 seats) under AV rules. The Conservative Party was the only party to have a lower seat count under AV rules in this election. The major losses in this election for the Conservatives came in Ontario, British Co-lumbia, and Saskatchewan, where the party lost 11, 6, and 6 seats, respec-tively. Strong NDP second-choice voter preferences for the Liberals over the Conservatives would have helped the Liberals take seats away from the Conservatives, who only averaged 38 percent of the vote in the ridings that they had initially won under FPTP, but lost under AV.

The 2006 election is our only case where AV voting rules may have changed the party in power (table 6). Using AV, the seat total for the Con-servatives was lowered by 36, reducing their share of seats in the House of Commons to 29 percent from 40 percent. The Liberals, on the other hand, gained 22 seats, pushing their share of seats up to 41 percent from 33 percent. The result of these swings was a change of government from a minority Conservative government to a minority Liberal government. Indeed, the Conservatives were the only party to lose any seats when AV


Canadian Alliance

PC Green Liberal NDP Bloc

FPTP AV FPTP AV FPTP AV FPTP AV FPTP AV FPTP AV

Total Seats

Won

66 58 12 9 0 0 172 187 13 8 38 39

% of Seats

Won

22% 19% 4% 3% 0% 0% 57% 62% 4% 3% 13% 13%





was applied to the 2006 election, as both the BQ and NDP saw gains. The Conservatives had their greatest loses in Ontario, where they were esti-mated to have 20 fewer seats under AV. All of the Conservative Party loses in Ontario went to the Liberal Party. The Conservatives did not fare well in Saskatchewan either, losing six seats in the province. All of these seats went to the NDP, which has a long history in Saskatchewan. One interest-ing change in parliamentary dynamics that would have occurred with the hypothetical implementation of AV is that the position of the Liberals, as well as the NDP and BQ, would have increased enough that the Liberals would have needed only the support of either the NDP or BQ in order to govern with a majority coalition. Such an outcome could have had an in-teresting effect on policies, since the Liberals would likely have been pulled either to the left, or towards implementing policies that were Quebec-cen-tric, in order to get the support of one of the two parties to form a coalition. When the Conservatives held the minority government under the FPTP rules, the Liberals would have needed the support of both the NDP and the BQ to form a coalition.

The application of AV voting rules to the 2008 election resulted in no change to the outcome (table 7). Under FPTP, the Conservatives had a minority government, and that would still have been the case under an AV system. There are a few different reasons why the Conservatives would


Conservative Green Liberal NDP Bloc Independent


Total Seats

Won99 71 0 0 135 160 19 19 54 57 1 1

% of Seats

Won32% 23% 0% 0% 44% 52% 6% 6% 18% 19% 0% 0%

Result FPTP: Liberal Minority




have kept their highest seat count. Although the Conservatives lost a total of 21 seats under the AV system, they simply did not lose enough seats to the Liberals for the Liberals to surpass their total seat count. Part of the reason for this is that the Conservatives performed much better than they had in previous elections, which would have lost them only six seats to the Liberals this time around under an AV system. This compares with the 20 seats that the Conservatives would have lost to the Liberals in the previous election. Indeed, in the 2006 election when the Conservatives lost a large number of seats in Ontario, their average vote share in Ontario constitu-encies was 34 percent, compared with 38 percent in 2008. Another reason that AV voting rules did not change the outcome of the 2008 election is that the Liberals faced a greater challenge from the left than in previous elections based on the performance of the NDP, particularly in Ontario, which is where the Liberals had typically gained their most seats from the Conservatives when elections were estimated with AV rules.

The 2011 election is the election of the so called “orange crush,” when the NDP surged into the official opposition after both the Liberals and the BQ saw their support collapse (table 8). This election is another one where AV voting rules could have had an impact on the outcome. Under FPTP, the outcome was a Conservative majority, with the Conservatives controlling 54 percent of the seats in the House of Commons. This compared to the opposition NDP, who controlled 33 percent of the seats, while the Liberals




Total Seats

Won124 88 0 0 103 125 29 40 51 54 1 1

% of Seats

Won40% 29% 0% 0% 33% 41% 9% 13% 17% 18% 0% 0%

Result FPTP: Conservative Minority

Result AV: Liberal Minority



held only 11 percent of the seats. When the election outcomes were esti-mated using AV voting rules, the Conservatives would have lost enough seats to see their share of the total drop to 44 percent. Both the NDP and Liberals would have seen their share of seats rise. The NDP would have had a large gain of 19 seats, giving them an overall share of 40 percent and placing them very close to the Conservative total. The success of the NDP was mostly the result of the poor performance by the Liberals across the country. In cases where the Liberal candidate would have been eliminated, 63 percent of their votes, on average, would have gone to the NDP, com-pared with only 21 percent going to the Conservative candidate. What also contributed to the large loss of 29 seats by the Conservatives under an AV system was the Liberal gain of 14 seats in Ontario. The Liberals, however, would also have lost seats in Quebec to the NDP.

The final election analyzed using the AV simulation was the 2015 elec-tion in which the Liberals came back to power for the first time in almost 10 years. AV electoral rules in this election would not have changed the governing party. Had AV been in place, the Liberals likely would have seen their position strengthen by 31 seats, giving them a total of 215 seats or 64 percent of the seats in the legislature.7 As with previous elections, much

7 Our results differ somewhat from Grenier (2015a, 2015b), who also simulated the 2015 election using AV rules. Grenier estimated that AV would have resulted in 224 seats




Total Seats

Won143 122 0 0 77 88 37 43 49 54 2 1

% of Seats

Won46% 40% 0% 0% 25% 29% 12% 14% 16% 18% 1% 0%

Result FPTP: Conservative Minority

Result AV: Conservative Minority



of this gain for the Liberals would have come at the expense of the Con-servatives, particularly in Ontario. Ontario is where the Liberals would have gained 19 seats, while the Conservatives lost 21. The Liberals would have also seen large gains in British Columbia. In this province the Liberals gained 11 seats—9 from the Conservatives and 2 from the NDP.

While having AV electoral rules during the 2015 election would not have changed the party that holds power, it would have likely lead to a shift in the party holding Official Opposition status (table 9). The NDPs gain of 15 seats and the Conservative’s loss of 42 would have led these two parties to switch positions in the legislature, as the NDP would have had a higher seat share than the Conservatives, making them the Official Opposition. Almost all of the NDPs gains came in Quebec, where they were able to take seats away from the Conservatives, Liberals, and the BQ.

for the Liberals and 50 for the NDP, where we have estimated 215 seats for the Liberals and 59 for the NDP. These differences were the greatest between the two estimates, and they appear to be attributable to differences in the effect of AV in Quebec. Our results for Quebec were more favorable to the NDP than the Liberals. Grenier estimated that the Liberals would have won 51 seats in Quebec, whereas we estimate that they would have received 36 seats. The seats that we estimated differently go to the NDP, for the most part, as our estimates indicate that the NDP would have won 29 seats in Quebec, while Grenier estimated that they would have received 20. These differences are likely the result of using different second-choice polling data. While the final number of seats distributed to each party differs between our analysis and Grenier’s, the overall result is the same: the Liberals would have further strengthened their majority under an AV system.


Conservative Green Liberal NDP Bloc

FPTP AV FPTP AV FPTP AV FPTP AV FPTP AV

Total Seats Won 166 137 1 1 34 47 103 122 4 1

% of Seats Won 54% 44% 0% 0% 11% 15% 33% 40% 1% 0%

Result FPTP: Conservative Majority

Result AV: Conservative Minority



The simulation of the previous seven elections using AV electoral rules has yielded some interesting results about how this type of voting system could change the dynamics between the parties. For instance, as figure 1 shows, in all but one case (the PCs in 1997), the Conservatives lost seats in every election under an AV system. Since the 2004 election, when the cur-rent Conservative party came into existence, the average number of seats lost by the Tories would have been approximately 31. Further, the Conser-vatives never would have been able to achieve a majority government.

The situation shows a much different outcome for the Liberal Party. In all of the seven elections that were part of this sample, the Liberals would have increased their number of seats under AV voting (figure 2). On aver-age, the Liberals would have improved their position by approximately 19 seats. Moreover, AV voting rules could have also resulted in changed out-comes in the 2004 and 2006 elections that favoured the Liberals. That the Liberals have the most to gain with AV voting are not in dispute and have been noted in other research. In his re-estimation of the 1997 election with AV electoral rules, Bilodeau’s (1999) analysis came to similar conclusions as ours: that the Liberals would have gained the largest number of seats of any party in the election, giving them an even stronger majority govern-ment. In their quasi-experiment of AV in Ontario during the province’s 2011 election, Blais et al. (2012) found that AV would have increased the number of seats won by the provincial Liberals, leading the authors to con-clude that AV tends to favour parties that tend to be more centrist. With


Conservative Green Liberal NDP Bloc


Total Seats Won 99 57 1 1 184 215 44 59 10 6

% of Seats Won 29% 17% 0% 0% 54% 64% 13% 17% 3% 2%





Figure 2: Performance of the Liberal Party Under FPTP and AV, 1997-2015

Figure 1: Performance of the Conservative Party Under FPTP and AV, 1997-2015

0

20

40

60

80

100

120

140

160

180

Reform PC Can.Alliance

PC Conserv. Conserv. Conserv. Conserv. Conserv.

1997 2000 2004 2006 2008 2011 2015

FPTP

AV

Num

bero

f sea

ts

0

50

100

150

200

250

1997 2000 2004 2006 2008 2011 2015

FPTP

AV

Num

bero

f sea

ts



the Liberals positioned between the Conservatives and NDP, this helps explain why AV elections would likely be biased in their favour.

The situation for the NDP under AV voting is closer to that of the Liber-als than that of the Conservatives. As figure 3 illustrates, in every election since 2006 the NDP gained an average of just under 13 seats by having AV voting rules in place instead of FPTP. It is likely that AV also would have led to the NDP being the Official Opposition in 2015 instead of the Con-servatives.

Conclusion and discussion

It is clear from the foregoing analysis that in the Canadian context, an AV electoral system favours the Liberals, and to a lesser extent, the NDP. Not only would the Liberal majority have been strengthened in the 2015 elec-

Figure 3: Performance of the New Democratic Party Under FPTP and AV, 1997-2015

0

20

40

60

80

100

120

140

1997 2000 2004 2006 2008 2011 2015

FPTP

AV

Num

bero

f sea

ts



tion, but as we have seen, they would have increased their electoral seats in all elections since 1997 using AV over FPTP. In some instances, the voting system would have resulted in different governments, and in other cases, different official opposition parties. Had AV been in place in earlier elec-tions, specifically the 2006 election, political and strategic considerations of the parties undoubtedly would have changed. Liberal leader Paul Martin may have stayed in Parliament, while in all likelihood Conservative leader Stephen Harper would have resigned.

Parties are adept at changing strategy in response to the rules of the game. The purpose of this exercise is not to demonstrate that AV works in the same way in every election, but to see whether AV produces similar ef-fects in different elections. As Ken Carty’s quote aptly states earlier in this chapter, changing the rules changes everything. However, what this exer-cise demonstrates is that of the current parties represented in the House of Commons, the Liberals are the party that has the least work to do in order to maintain power under an Alternative Voting system.

That conclusion should not give the Liberals false hope. As Jensen found in his study of Western Canadian provinces that switched to AV for their elections, the expected outcome is not the actual outcome. The elec-torate may see through the attempts by parties to grab power by shutting out other parties.

Even if the Liberals had an advantage under the AV system, that advan-tage does not preclude other parties from changing their strategies and approaches. For example, when the Liberal government of Jean Chretien changed campaign financing regulations in 2004 to limit large donations, the Liberals themselves suffered because they had previously focused on large donations and did not have a grassroots system of collecting smaller donations. In contrast, the Conservatives benefitted the most, not only be-cause they had a good method for acquiring small donations, but because they were competent at building their get-out-the-vote database and orga-nization. Any number of small changes to party strategy by the opposition parties could deliver unanticipated outcomes.

While it is clear that AV, given the current party system, would benefit the Liberal Party of Canada, it is less clear whether AV fulfills any of the



guiding principles set out by the government for the committee on elec-toral reform.

Because AV falls under the same majority/plurality electoral family as FPTP, rather than reduce distortions between voter intention and the elec-toral result, it actually amplifies those distortions when voters choose the Liberals as their second-choice selections. AV thus manufactures a ma-jority where none existed previously. Moreover, the fact that the Liberals would likely be the largest beneficiary under AV leads to questions as to how a switch to this system would restore legitimacy to the voting process.

While our analysis did not measure voter participation, it remains un-clear how changing to an AV system would encourage greater engagement and participation. Given the increased onus on voters to rank order their choices, that complexity could result in lower voter turnout—or it might have no effect at all. In addition, even if voter turnout was to be higher under AV, the expected higher rate of rejected ballots would offset some of this effect.

The third guiding principle for the committee of “avoiding undue com-plexity in the voting process” is partly addressed in that the change to an AV system only requires individuals mark their ballot with numbers rather than an X. However, as we found in prior attempts to use this system, if the legislation requires a rank ordering, or does not accept the traditional X, there is a high probability that the system will become more complex than the current one. While the voting itself need not suffer from greater complexity, the counting of the votes certainly would add complications.

AV does nothing to address the requirement of safeguarding the in-tegrity of the voting process, nor does it change the accountability of the local representative. In sum, the main outcome of moving to an AV system would be to strengthen the Liberal Party of Canada’s control and hold on the federal government. It would do nothing to fix the so-called problems of FPTP.



Appendix

The following tables provide the detailed results by province and party for the elections that were re-estimated using AV electoral rules.

ww

w.fraserinstitute.org d Fraser Institute


Table A1: Detailed 1997 Results Comparison

Prov. Reform PC Green Liberal NDP Bloc Independent


Alberta 24 23 0 0 0 0 2 3 0 0 0 0

British Columbia 25 19 0 0 0 0 6 13 3 2 0 0

Manitoba 3 0 1 3 0 0 6 6 4 5 0 0

New Brunswick 0 0 5 6 0 0 3 2 2 2 0 0

Newfoundland

and Labrador

0 0 3 4 0 0 4 3 0 0 0 0

Nova Scotia 0 0 5 5 0 0 0 0 6 6 0 0

Nunavut 0 0 0 0 0 0 1 1 0 0 0 0

Northwest

Territories

0 0 0 0 0 0 1 1 0 0 0 0

Ontario 0 0 1 1 0 0 101 101 0 0 1 1

Prince Edward Is. 0 0 0 0 0 0 4 4 0 0 0 0

Quebec 0 0 5 7 0 0 26 33 0 0 44 35 0 0

Saskatchewan 8 7 0 0 0 0 1 2 5 5 0 0

Yukon 0 0 0 0 0 0 0 1 1 0 0 0


Total Seats Won 60 49 20 26 0 0 155 170 21 20 44 35 1 1

% of Seats Won 19.9% 16.3% 6.6% 8.6% 0.0% 0.0% 51.5% 56.5% 7.0% 6.6% 14.6% 11.6% 0.3% 0.3%

Fraser Institute d ww

w.fraserinstitute.org



Prov. Canadian Alliance PC Green Liberal NDP Bloc


Alberta 23 23 1 1 0 0 2 2 0 0

British Columbia 27 25 0 0 0 0 5 8 2 1

Manitoba 4 5 1 0 0 0 5 7 4 2

New Brunswick 0 0 3 2 0 0 6 7 1 1

Newfoundland

and Labrador0 0 2 2 0 0 5 5 0 0

Nova Scotia 0 0 4 4 0 0 4 7 3 0

Nunavut 0 0 0 0 0 0 1 1 0 0

Northwest Territories 0 0 0 0 0 0 1 1 0 0

Ontario 2 0 0 0 0 0 100 103 1 0

Prince Edward Is. 0 0 0 0 0 0 4 4 0 0

Quebec 0 0 1 0 0 0 36 36 0 0 38 39

Saskatchewan 10 5 0 0 0 0 2 5 2 4

Yukon 0 0 0 0 0 0 1 1 0 0


Total Seats Won 66 58 12 9 0 0 172 187 13 8 38 39

% of Seats Won 21.9% 19.3% 4.0% 3.0% 0.0% 0.0% 57.1% 62.1% 4.3% 2.7% 12.6% 13.0%

ww




Prov. Conservative Green Liberal NDP Bloc Independent


Alberta 26 25 0 0 2 3 0 0 0 0


Manitoba 7 5 0 0 3 6 4 3 0 0

New Brunswick 2 2 0 0 7 7 1 1 0 0

Newfoundland

and Labrador

2 0 0 0 5 7 0 0 0 0

Nova Scotia 3 2 0 0 6 8 2 1 0 0

Nunavut 0 0 0 0 1 1 0 0 0 0


Ontario 24 13 0 0 75 88 7 5 0 0

Prince Edward Is. 0 0 0 0 4 4 0 0 0 0

Quebec 0 1 0 0 21 17 0 0 54 57 0 0

Saskatchewan 13 7 0 0 1 4 0 3 0 0

Yukon 0 0 0 0 1 1 0 0 0 0


Total Seats Won 99 71 0 0 135 160 19 19 54 57 1 1

% of Seats Won 32.1% 23.1% 0.0% 0.0% 43.8% 51.9% 6.2% 6.2% 17.5% 18.5% 0.3% 0.3%







Alberta 28 28 0 0 0 0 0 0 0 0


Manitoba 8 5 0 0 3 6 3 3 0 0

New Brunswick 3 2 0 0 6 7 1 1 0 0

Newfoundland

and Labrador

3 3 0 0 4 4 0 0 0 0

Nova Scotia 3 1 0 0 6 5 2 5 0 0

Nunavut 0 0 0 0 1 1 0 0 0 0


Ontario 40 20 0 0 54 74 12 12 0 0

Prince Edward Is. 0 0 0 0 4 4 0 0 0 0

Quebec 10 10 0 0 13 10 0 0 51 54 1 1

Saskatchewan 12 6 0 0 2 2 0 6 0 0

Yukon 0 0 0 0 1 1 0 0 0 0


Total Seats Won 124 88 0 0 103 125 29 40 51 54 1 1

% of Seats Won 40.3% 28.6% 0.0% 0.0% 33.4% 40.6% 9.4% 13.0% 16.6% 17.5% 0.3% 0.3%

ww






Alberta 27 27 0 0 0 0 1 1 0 0


Manitoba 9 9 0 0 1 1 4 4 0 0

New Brunswick 6 3 0 0 3 6 1 1 0 0

Newfoundland

and Labrador

0 0 0 0 6 6 1 1 0 0

Nova Scotia 3 2 0 0 5 3 2 5 1 1

Nunavut 1 0 0 0 0 1 0 0 0 0


Ontario 51 45 0 0 38 44 17 17 0 0

Prince Edward Is. 1 0 0 0 3 4 0 0 0 0

Quebec 10 7 0 0 14 13 1 1 49 54 1 0

Saskatchewan 13 11 0 0 1 1 0 2 0 0

Yukon 0 0 0 0 1 1 0 0 0 0


Total Seats Won 143 122 0 0 77 88 37 43 49 54 2 1

% of Seats Won 46.4% 39.6% 0.0% 0.0% 25.0% 28.6% 12.0% 14.0% 15.9% 17.5% 0.6% 0.3%





Prov. Conservative Green Liberal NDP Bloc


Alberta 27 27 0 0 0 0 1 1

British Columbia 21 19 1 1 2 2 12 14

Manitoba 11 9 0 0 1 1 2 4

New Brunswick 8 6 0 0 1 1 1 3

Newfoundland and Labrador 1 0 0 0 4 5 2 2

Nova Scotia 4 4 0 0 4 4 3 3

Nunavut 1 1 0 0 0 0 0 0

Northwest Territories 0 0 0 0 0 0 1 1

Ontario 73 57 0 0 11 25 22 24

Prince Edward Island 1 1 0 0 3 3 0 0

Quebec 5 3 0 0 7 4 59 67 4 1

Saskatchewan 13 10 0 0 1 1 0 3

Yukon 1 0 0 0 0 1 0 0


Total Seats Won 166 137 1 1 34 47 103 122 4 1

% of Seats Won 53.9% 44.5% 0.3% 0.3% 11.0% 15.3% 33.4% 39.6% 1.3% 0.3%

ww




Prov. Conservative Green Liberal NDP Bloc


Alberta 29 26 0 0 4 6 1 2

British Columbia 10 1 1 1 17 28 14 12

Manitoba 5 4 0 0 7 9 2 1

New Brunswick 0 0 0 0 10 10 0 0

Newfoundland and Labrador 0 0 0 0 6 6 0 0

Nova Scotia 0 0 0 0 11 11 0 0

Nunavut 0 0 0 0 1 1 0 0

Northwest Territories 0 0 0 0 1 1 0 0

Ontario 33 12 0 0 81 100 8 10

Prince Edward Island 0 0 0 0 4 4 0 0

Quebec 12 7 0 0 40 36 16 29 10 6

Saskatchewan 10 7 0 0 1 2 3 5

Yukon 0 0 0 0 1 1 0 0


Total Seats Won 99 57 1 1 184 215 44 59 10 6

% of Seats Won 29.3% 16.9% 0.3% 0.3% 54.4% 63.6% 13.0% 17.5% 3.0% 1.8%



References

Australian Electoral Commission (2015). Informality (%) House of Representatives and Senate. Web table. Government of Australia. <http://www.aec.gov.au/Voting/Informal_Voting/summary.htm>, as of August 24, 2016.

Bean, Clive (1997). Australia’s Experience with the Alternative Vote. Representation 34, 2: 103-110.

Bilodeau, Antonie (1999). L’impact mécanique du vote alternatif au Canada: une simulation des élections de 1997. Canadian Journal of Political Science 32, 4 (December): 745-761.

Blais, André, Maxime Héroux-Legault, Laura Stephenson, William Cross, and Elisabeth Gidengil (2012). Assessing the Psychological and Mechanical Impact of Electoral Rules: A Quasi-Experiment. Electoral Studies 31: 829-837.

Canada (n.d.). Democratic Institutions: Frequently Asked Questions. Web page. Government of Canada. <http://www.democraticinstitutions.gc.ca/eng/content/frequently-asked-questions-er>, as of September 27, 2016.

Canada (2016). Frequently Asked Questions (FAQ) about Canadian Federal Electoral Reform. Web page. Government of Canada. <https://www.canada.ca/en/campaign/electoral-reform/learn-about-canadian-federal-electoral-reform/frequently-asked-questions-about-canadian-federal-electoral-reform.html>, as of September 27, 2016.

Carty, Kenneth. (2016). House of Commons Special Committee on Electoral Reform. Evidence: Monday, July 25, 2016. Parliament of Canada. <http://www.parl.gc.ca/HousePublications/Publication.aspx?Language=e&Mode=1&Parl=42&Ses=1&DocId=8392670#Int-9017257>, as of October 11, 2016.

https://www.canada.ca/en/campaign/electoral-reform/learn-about-canadian-federal-electoral-reform/frequently-asked-questions-about-canadian-federal-electoral-reform.html






Clemens, Jason, Taylor Jackson, Steve LaFleur, and Joel Emes (2016). Electoral Rules and Fiscal Policy Outcomes. In Lydia Miljan (ed.). Counting Votes: Essays on Electoral Reform. Fraser Institute.

Elections Canada (2015). Official Voting Results: Forty-Second General Election. Elections Canada. <http://www.elections.ca/res/rep/off/ovr2015app/home.html#1>, as of August 24, 2016.

Elkins, David, J. (1976). Politics Makes Strange Bedfellows: The B.C. Party System in the 1952 and 1953 Provincial Elections. B.C. Studies 30 (Summer): 3-26.

Fair Vote Canada (n.d.). Myth #6 – Is a “Ranked Ballot” a Voting System?? Web page. Fair Vote Canada. <http://www.fairvote.ca/ranked-ballot-is-not-a-voting-system/>, as of September 28, 2016.

Flanagan, Tom (1999). The Alternative Vote: An Electoral System for Canada. In Henry Milner (ed.). Making Every Vote Count: Reassessing Canada’s Electoral System (Broadview Press): 85-90.

Grenier, Éric (2015a, November 26). Change to Preferential Ballot would Benefit Liberals. CBC News. <http://www.cbc.ca/news/politics/grenier-preferential-ballot-1.3332566>, as of August 19, 2016.

Grenier, Éric (2015b). Riding Results with a Preferential Ballot. Web blog. ThreeHundredEight.com. <http://www.threehundredeight.com/2015/11/riding-results-with-preferential-ballot.html>, as of August 19, 2016.

Jansen, Harold, J. (2004). The Political Consequences of the Alternative Vote: Lessons from Western Canada. Canadian Journal of Political Science 37, 3 (September): 647-669.

http://www.fairvote.ca/ranked-ballot-is-not-a-voting-system/

http://www.fairvote.ca/ranked-ballot-is-not-a-voting-system/



Jansen, Harold, J. (2015, November 30). Reforming the Electoral System. Policy Options. <http://policyoptions.irpp.org/magazines/november-2015/reforming-the-electoral-system/>, as of October 7, 2016.

Kay, Barry (2016, June 2). Electoral Reform – Why We Should Prefer the Preferential Ballot. Ottawa Citizen. <http://ottawacitizen.com/opinion/columnists/kay-electoral-reform-why-we-should-prefer-the-preferential-ballot>, as of October 7, 2016.

Lijphart, Arend. (1994). Electoral Systems and Party Systems: A Study of Twenty-Seven Democracies 1945-1990. Oxford: Oxford University Press.

Milner, Henry (ed). (1999). Making Every Vote Count: Reassessing Canada’s Electoral System. Broadview.

Northrup, David (2012). The 2011 Canadian Election Survey: Technical Documentation. The Canadian Election Study (CES). Canadian Opinion Research Archive. <http://www.queensu.ca/cora/ces.html>, as of August 11, 2016.

Pepall, John (2016). First-Past-the-Post: Empowered Voters, Accountable Government. In Lydia Miljan (ed.). Counting Votes: Essays on Electoral Reform. Fraser Institute.

Rydon, Joan (1956). Electoral Methods and the Australian Party System, 1910-1951. Australian Journal of Politics & History 2, 1: 68-83.

Sharman, C., A.M. Sayers, and N. Miragliotta (2002). Trading Party Preferences: The Australian Experience of Preferential Voting. Electoral Studies 21: 543-560.

http://policyoptions.irpp.org/magazines/november-2015/reforming-the-electoral-system/

http://policyoptions.irpp.org/magazines/november-2015/reforming-the-electoral-system/

http://ottawacitizen.com/opinion/columnists/kay-electoral-reform-why-we-should-prefer-the-preferential-ballot





Data

Blais, André, Elisabeth Gidengil, Neil Nevitte, Patrick Fournier, Joanna Everitt (2006). The 2006 Canadian Election Study. Canadian Opinion Research Archive. <http://www.queensu.ca/cora/ces.html>, as of August 17, 2016.

Blais, André, Elisabeth Gidengil, Neil Nevitte, Patrick Fournier, Joanna Everitt (2004). The 2004 Canadian Election Study. Canadian Opinion Research Archive. <http://www.queensu.ca/cora/ces.html>, as of August 17, 2016.

Blais, André, Elisabeth Gidengil, Richard Nadeau, Neil Nevitte (2000). The 2000 Canadian Election Study. Canadian Opinion Research Archive. <http://www.queensu.ca/cora/ces.html>, as of August 17, 2016.

Blais, André, Elisabeth Gidengil, Richard Nadeau, Neil Nevitte (1997). The 1997 Canadian Election Study. Canadian Opinion Research Archive. <http://www.queensu.ca/cora/ces.html>, as of August 17, 2016.

Fournier, Patrick, Fred Cutler, Stuart Soroka, and Dietlind Stolle (2016). The 2015 Canadian Election Study. Canadian Opinion Research Archive. <http://ces-eec.arts.ubc.ca/english-section/surveys/>, as of August 17, 2016.

Fournier, Patrick, Fred Cutler, Stuart Soroka and Dietlind Stolle (2011). The 2011 Canadian Election Study. Canadian Opinion Research Archive. <http://www.queensu.ca/cora/ces.html>, as of August 17, 2016.

Gidengil, Elisabeth, Joanna Everitt, Patrick Fournier, Neil Nevitte (2008). The 2008 Canadian Election Study. Canadian Opinion Research Archive. <http://www.queensu.ca/cora/ces.html>, as of August 17, 2016.

Kollman, Ken, Allen Hicken, Daniele Caramani, David Backer, and David Lublin (2016). Constituency-Level Elections Archive. Center for Political Studies, University of Michigan. <http://www.electiondataarchive.org/index.html>, as of August 11, 2016.



About the authors

Lydia Miljan is an Associate Professor of Political Science at the University of

Windsor, and a Fraser Institute Senior Fellow. She was Director of the Fraser

Institute’s Alberta Policy Research Centre and from 1988 to 2001 was the Chair of the

National Media Archive. She holds a Ph.D. in Political Science from the University

of Calgary, specializing in political communication. Prof. Miljan teaches in the areas

of Canadian public policy, research methodology, and politics and the media. Her

main research interests include how journalist’s personal views are reflected in

news content and public opinion formation. In addition to peer-reviewed papers,

she is the author of three books: Public Policy in Canada, Hidden Agendas: How

Journalists Influence the News, and Cross-Media Ownership and Democratic Practice

in Canada. Hidden Agendas was short-listed for the Donner Prize for the best book

in public policy, 2003/04.

Taylor Jackson is a Senior Policy Analyst at the Fraser Institute. He holds a B.A.

and an M.A. in Political Science from Simon Fraser University. Mr. Jackson is the

co-author of a number of studies and book chapters on a wide range of policy issues,

including energy, mining and natural resources, democratic reform, Canadian-

American relations, and other areas. Mr. Jackson’s work has been covered in the

media around the world and his commentaries have appeared in the National Post,

Financial Post, and Washington Times, as well as other newspapers across Canada.


CHAPTER 4 The Imperative of a ReferendumPatrice Dutil 1

The throne speech delivered to the Canadian Parliament by the new Lib-eral government in December 2015 made one very specific promise, which was crystal clear both in its commitment and its timing. “To make sure that every vote counts,” it declared, “the government will undertake con-sultations on electoral reform, and will take action to ensure that 2015 will be the last federal election conducted under the first-past-the-post voting system.”

The implications of this declaration are heavy for Canada’s democracy and for its political future. The obvious question has to be addressed: What is the point of promising consultations if the government has already made up its mind to proceed with making fundamental changes? The govern-ment of Canada must indeed consult widely, but more than that, it must put its reform proposals to the people of Canada in a referendum. It has no mandate to push its reforms through by some arbitrary, hard deadline.

1 Patrice Dutil is Professor in the Department of Politics and Public Administration at Ryerson University in Toronto. He has worked as a government policy advisor, a non-profit organization executive, a television producer, and a magazine editor. He is currently the President of the Champlain Society (www.champlainsociety.ca) and the editor of Canadian Government Executive magazine. He was the founder and edi-tor (1991-1996) of The Literary Review of Canada (www.reviewcanada.ca). His main research interests are political and public sector leadership, institutionalism, governance, and the policy development process. He has published seven books, and regularly con-tributes to scholarly journals. His website is http://patricedutil.com.

http://www.champlainsociety.ca/

http://www.reviewcanada.ca/

http://patricedutil.com



On four separate occasions, Canadians from various parts of the country expressed their support for the first-past-the-post voting system. The gov-ernment of Canada cannot impose electoral procedures that have been rejected without asking for the right to do so. Such an imposition would be simply undemocratic, an act outside the conventions and precedents of this country.

Since the first election in Canada in 1792, Canadians have used the simple plurality system (commonly known as the “first-past-the-post” sys-tem) to elect their representatives to the House of Commons, the provin-cial assemblies, and municipal governments. It remains a conventional practice used in most Canadian civic, corporate, and social situations that require a vote to choose representatives.

For 225 years, it has been a broadly accepted practice in Canada, part of the very fabric of the country’s political culture. There has been no clamour for it to be changed; the demands that it be changed in time for the 2019 election must have been expressed privately. Furthermore, while the Lib-eral Party inserted a few lines in its election manifesto about changing the electoral system because it found it “undemocratic,” it won a dispropor-tionate majority under those very same rules. The election it won, with almost 60% of the electorate voting against it, did not give it a mandate to transform the electoral rules that have been rooted in this country’s con-stitutional conventions for more than ten generations.

This chapter examines the nature of the precedents that have been set in Canada and in other Westminster systems over the past decades for the use of plebiscites to effect electoral reform. The record is clear. Before any changes to Canada’s long-established practices are implemented, political leaders have asked voters for their assent. As such, they have established an unavoidable convention.

The duty to consultPrecedents and conventions matter.

In 1980, the newly elected government (elected with 44% of the vote) headed by Pierre Elliott Trudeau launched a drive to patriate the Canadian


The Imperative of a Referendum d 83

Constitution and to amend it with a Charter of Rights and Freedoms. It drafted a Proposed Resolution for a Joint Address to Her Majesty the Queen respecting the Constitution of Canada that, presuming to speak on be-half of Canadians, laid out a series of amendments. A number of prov-inces challenged this procedure. Their argument hinged on the fact that on many occasions, the federal government had come to an agreement with the provinces before petitioning Westminster to draft an amendment to the British North America Act (the BNA Act). Canada, they argued, had been established for the benefit of the provinces; it was inconceivable that it could be changed against their explicit wishes. The BNA Act’s preamble was clear that “the Provinces of Canada, Nova Scotia, and New Brunswick have expressed their Desire to be federally united into One Dominion un-der the Crown of the United Kingdom of Great Britain and Ireland, with a Constitution similar in Principle to that of the United Kingdom” and that, in particular, “such a Union would conduce to the Welfare of the Prov-inces and promote the Interests of the British Empire.” For the provinces, it was imperative: they had to approve any constitutional amendment that affected their rights.

That idea was rejected by Pierre Trudeau’s Liberal government. It ar-gued that it had a clear mandate from the people and the constitution did not explicitly require the consent of provinces by virtue of the fact that there was no amending formula in the BNA Act. The government’s po-sition hinged on the assumption that Canada was a constitutional being that could act on its own. Manitoba, Newfoundland, and Quebec acted on their convictions that Ottawa could not act unilaterally and each took their cases to their respective Courts of Appeal.

Manitoba put three theoretical questions to its highest court. The first asked if the amendment to the BNA Act proposed by Ottawa would af-fect the rapport between the national government and the provinces as well as the very powers enjoyed by their governments. The second ques-tion probed the issue of constitutional understanding: was it not a matter of constitutional convention that any change to the relationships between the federal government and the provinces—or indeed their very powers—could not be made without the agreement of the provinces? The third and



final question plainly asked if the BNA Act implicitly required that any amendment had to have the agreement of the provinces.

The Manitoba Court found that indeed the government of Canada’s proposal would affect provincial powers, but responded negatively to the last two questions. The matter of constitutional convention was too murky and it found nothing in the constitution that prevented Ottawa from act-ing unilaterally, without the consent of the provinces.

Newfoundland posed the same questions as Manitoba, but also added a question (mindful that a referendum had sanctioned Newfoundland’s union with Canada in 1949). It asked: “Could the Terms of Union between Newfoundland and Canada be amended directly or indirectly… without the consent of the Government, Legislature or a majority of the people in the Province of Newfoundland voting in a referendum?” (SCC Re: Resolu-tion to amend the Constitution: 763). The Newfoundland Court of Appeal found that the government of Canada could not act unilaterally.

Quebec also asked its highest court essentially the same questions, and received similar answers. The Quebec Court of Appeal responded that, in its judgment, the federal government’s proposed amendments would in-deed have an impact on the province’s competence and would even have an impact on the status of the provincial legislatures within the Canadian federation. It also held that Ottawa did have the authority to make unilat-eral changes to the Constitution.

Unsatisfied by their own courts’ answers, the three provinces appealed to the Supreme Court of Canada, which heard all three appeals together. In SCC Re: Resolution to amend the Constitution, the Court unanimous-ly ruled that it agreed with the responses of the Manitoba and Quebec Courts of Appeal to the first questions. There was no doubt that Ottawa’s proposed changes to the Constitution would indeed affect the “powers, rights or privileges” of the provinces.

The Court also examined the two other questions: the issue of the le-gality of Ottawa’s plans, and the issue of whether a convention existed. A majority of the Court judged that the federal government did in fact have the legal authority to unilaterally approach the United Kingdom’s Parlia-ment for a constitutional amendment.



In its third ruling, the Court also affirmed the existence of constitu-tional conventions in Canada, and a majority of the justices found that the government of Canada’s plan did in fact violate the conventions by try-ing to act unilaterally. However, it shied away from imposing itself further, stating essentially that it was not up to the courts to uphold conventions—its duty was to ensure that formal rules were obeyed. In other words, there were certainly precedents and conventions that had to be observed, but these were political, not legal, and thus beyond the purview of jurispru-dence. “The requirements for establishing a convention bear some resem-blance with those which apply to customary law. Precedents and usage are neces sary but do not suffice,” the court stated (SCC Re: Resolution to amend the Constitution: 888).

The key point from the Supreme Court of Canada’s statement was that precedents and conventions mattered. They are important because they capture a certain idea of political culture and practice. In the context of the British system, one which works without a constitutional text and is therefore instructed by past actions, the British expert Sir Ivor Jennings argued that constitutional conventions “provide the flesh that clothes the dry bones of the law; they make the legal constitution work; they keep it in touch with the growth of ideas” (Jennings, 1959: 136). Jennings articulated a series of ques-tions to test the validity of a convention. For him, three conditions had to be met in order to do so and together they became known at the “Jennings Test.”

• Were there precedents?• Did the key actors in the precedents believe that they were bound

by a rule?• Would there be a constitutional reason for the rule?

These were the considerations that the Supreme Court of Canada ap-plied to gauge the resilience of the convention and whether it had with-stood a test of time. For Jennings, all three questions had to be satisfac-torily answered in order to judge a practice a valid “convention.” In giving answers to the reference put to it by Pierre Trudeau’s government, the Su-preme Court of Canada found that provincial consent had to be secured



because there were indeed a number of precedents. The Court pointed to five amendments (1930, 1931, 1940, 1951, and 1964) that directly affected federal-provincial relationships by changing provincial legislative powers. It observed that the provinces had been consulted and had agreed to the amendments. Past federal governments did believe that they had to se-cure provincial support in order for the actions to be ultimately legitimate. Not least, the constitutional division of powers between the government of Canada and the provinces lent support to the convention.

Consequently, it judged that there did exist a convention by which Ot-tawa, the provinces, and even the British Parliament had lived in order to change the constitution. The court summed up its judgment by saying that

“the agreement of the provinces of Canada, no views being expressed as to its quantification, is constitutional ly required for the passing of the Pro-posed Reso lution for a Joint Address to Her Majesty the Queen respecting the Constitution of Canada and that the passing of this Resolution with-out such agreement would be unconstitutional in the con ventional sense” (SCC Re: Resolution to amend the Constitution: 909). Instructed by the Su-preme Court that it needed a “substantial measure” of provincial consent, the Trudeau government redoubled its efforts and secured the approval of all the provinces, save Quebec, for its constitutional reform package.

The convention of plebiscites to reform voting practice in Canada

The voting system in practice in Canada is not enshrined in the constitu-tion. The Constitution Act does specify that members of Parliament must be “elected” but says nothing about what system is to be used to choose winners. There is, moreover, no constitutional amending formula that ap-plies to any changes in the way Canadians vote. However, there are prec-edents and conventions about how elections are determined have been part of the Canadian political culture for over a century. The Jennings test for conventions thus applies.

Over the past decade, five provincial governments—Prince Edward Island (2003), Ontario (2007), New Brunswick (2008) Quebec (2003-



2007), and British Columbia (2005 and 2009)—considered changing the voting system in their jurisdictions. Living up to the conventions of elections in Canada, four committed to put the question to the people: British Columbia (2005 and 2009), Prince Edward Island (2005), and Ontario (2007). In New Brunswick in 2006, the Progressive Conserva-tive government led by Premier Bernard Lord promised a plebiscite on electoral reform in 2008, but it was never held as the government was defeated. Quebec considered a number of proposals, but plans were quietly abandoned.

British Columbia, 2005 and 2009

In British Columbia, the Liberal party, led by Gordon Campbell, included a commitment to electoral reform in its 2001 election campaign. Its plan was to “Appoint a Citizens’ Assembly on Electoral Reform to assess all pos-sible models for electing the MLAs, including preferential ballots, propor-tional representation, and our current electoral system.” It also promised that the Citizens’ Assembly would hold public hearings throughout the province, and that “if it recommends changes to the current electoral sys-tem, that option will be put to a province-wide referendum” (British Co-lumbia Liberal Party, 2001: 30).

The Throne speech of 2003 entrenched the process. In April of that year, the government introduced a motion to establish the Citizen’s Assembly (British Columbia, 2003, 4th Session, 37th Parliament). It was unanimous-ly supported in the legislature (British Columbia, 2003, April 28).

At the same time, the government committed to holding a referendum two years later, in May 2005. Premier Campbell noted that: “This has never happened before, where a legislative body has said to the people who elect them that they want the advice of the public on how we should elect our elected representatives in the province of British Columbia. Indeed, if you go back to 1858, this is the first time in 145 years we are actually giving the people of British Columbia a direct say in how they should elect the MLAs that are meant to serve them. After all, in a democracy, we should



remember we are here at the service and the pleasure of the people of this province” (British Columbia, 2003, April 30).

He continued by emphasizing that “There is no more fundamental te-net that we agree to as we seek office than that in a democracy, the rules of the democracy should be designed by the people they serve, not by the power brokers who may wish that the democracy worked in their interests. It is by turning to the people and trusting the public that I believe we can re-establish the critical link between our democratic institutions and those that they are supposed to serve.”

Finally, he justified the need for a plebiscite by stating that “The govern-ment wants to ensure that all British Columbians have an opportunity to vote before any change is adopted. We want to be sure any change that is adopted is truly endorsed by the regions of the province and the people of the province” (British Columbia, 2003, April 30; British Columbia, Citi-zens’ Assembly on Electoral Reform, 2004).

The government established a minimum level of support for the plebi-scite results to be accepted. For reform to be enacted, at least 60% of the valid votes had to be cast in support of any proposal and a simple majority in favour in at least 60% of all electoral districts (48 out of 79) had to be achieved.2 Many argued that the threshold was too demanding. Premier Campbell defended the decision thus: “We believe this is a fundamental and significant change, and we therefore have placed a double approval process in place. There are some who have already suggested that that is too high an approval rating. Clearly, the government disagrees with that. We believe this is a significant change. It’s a significant change that should require the kind of approval that says, indeed, a great majority of people in this province feel that they will benefit from this change” (British Colum-bia, 2003, April 30).

2 It is worth noting that governments require a two-thirds vote of shareholders for fundamental changes or “special resolutions” to corporate statutes. For example, see s.1 of the Business Corporations Act (Ontario). I am indebted to Mr. Stephen Thiele for this observation.



The question put to the voters was clear: “Should British Columbia change to the BC-STV electoral system as recommended by the Citizens’ Assembly on Electoral Reform?” Nearly 58% of the citizens who cast a bal-lot supported the pro posed single transferable vote (STV) electoral system, a proportional representation method that allows voters to express their choices for office by order of preference in a multi-member constituency. A solid majority supported STV in all but two of BC’s 79 constituencies. But that was not enough to pass, according to the rules this legislature had unanimously established. The participation rate was relatively low on such an important issue: 58.2%. In effect, therefore, just over 33% of eligible vot-ers supported the initiative.

The proposal was rejected by voters, but the government renewed its commitment to electoral reform. In the Throne speech of September 2005, a few months after the first proposal for electoral reform was defeated in the plebiscite, the government declared that it “has been clear that it does not in tend to rewrite those rules after the fact or to pretend that the vote for STV succeeded when it did not. Nor can it ignore the size of the dou-ble majority that voted to change our current electoral system to the STV model” (British Columbia, 2005, September 12).

Just as significantly, the government stated that it did “not accept that the solution to a majority vote that failed to pass is to essentially ignore it and impose yet another electoral system. It does not accept that the answer to the minority’s rejec tion of the Citizens’ Assembly proposal is to redo its work. It does not accept that the 79 members of this assembly are any better qualified than the 161 mem bers of the Citizens’ Assembly were to choose the best electoral model” (British Columbia, 2005, September 12). The British Columbia government nevertheless committed itself to con-tinue to explore electoral reform. A new Electoral Boundaries Commis-sion was given the task of examining the provincial electoral map and then to examine and report back on the “the best and fairest way to configure British Columbia’s electoral dis tricts under the STV model.”

The government also again committed itself to seeking public support. The 2005 Speech from the Throne promised an “extensive effort to better inform British Co lumbians about the two electoral options: the current



system and STV” and that equal funding will be provided “to support ac-tive information campaigns for supporters and detractors of each model” (British Columbia, 2005, September 12: 28). It promised that the two mod-els would be put to a province-wide vote, along with the applicable elec-toral boundaries, in a referendum that would be held in tandem with the No vember 2008 municipal elections. The government also indicated that the premier would remain neutral and that the rules and thresholds that had applied for passing STV in the 2005 referendum would again apply.

The second plebiscite was eventually held during the provincial elec-tion of 12 May 2009. Voters were asked to vote on whether they preferred the STV method or the traditional “simple majority” system:

Which electoral system should British Columbia use to elect members to the provincial Legislative Assembly?• The existing electoral system (first-past-the-post)• The single transferable vote electoral system (BC-STV) proposed

by the Citizens’ Assembly on Electoral Reform

The STV proposal earned the support of 31% of voters—a dramatic drop in support compared with the first plebiscite—while 61% voted against it. The turnout for the election/plebiscite was 51%, indicating that popular support for the STV alternative was less than 10% of eligible voters.

Prince Edward Island, 2005 and 2015

In the 2000 election in Prince Edward Island that saw the re-election of the Progressive Conservative government under Premier Pat Binns, the opposition had gathered close to 45% of the vote, but was reduced to one representative in the legislature. In 2001, the province’s legislative assem-bly created a special committee to review the Elections Act and also com-missioned a study from Elections PEI to examine systems of proportional representation. That report was submitted to the legislature in April 2002. In January 2003, the legislature formed a Commission on Prince Edward



Island’s Electoral Future (Prince Edward Island, 2003, January 21). The commission concluded that a mixed member system would be appropri-ate (Prince Edward Island Electoral Reform Commission, 2003).

The government decided to put the matter to the voters of PEI. In mid-December 2004, Premier Binns stated that “We’re providing time for Islanders to debate this subject, to look at our current model, first past the post, to compare that to a mixed member system which would have some combination of first-past-the-post plus a slate, and to make a consid-ered judgment on which is the best and most appropriate model for PEI” (Prince Edward Island, 2004, December 16: 1036). Prince Edward Island’s House Speaker Gregory Deighan put it most eloquently: “It stands to rea-son,” he said, that Islanders “should have a strong voice in determining how these electoral systems work because they do have significant bearing on the final results of an election.” (Prince Edward Island, 2005, February 17). In the Throne Speech of November 2005, the government committed itself to holding a referendum in order to give “a significant opportunity for Is-landers to express their preference for our future electoral process” (Prince Edward Island, 2005, Nov. 16).

The referendum took place a few weeks later. The referendum ques-tion was: “Should Prince Edward Island change to the Mixed Member Proportional System as presented by the Commission of PEI’s Electoral Future?” In line with British Columbia’s practice, PEI also set a higher ceil-ing for approval. In order to pass, the referendum had to receive 60% of the province-wide popular vote and a simple majority in 60% (16 of 27) of the electoral districts. Voters rejected the proposed reform: 63.6% of eligible voters voted against the motion. Two districts (out of 27) carried a majority in favour of reform.

Almost a decade later, the June 2015 Throne speech committed the PEI government, now led by Liberal Premier Wade MacLauchlan, to another round of reflection on the province’s electoral system. A white paper on democratic renewal was submitted by the government to the legislature in the summer of 2015. The government also committed that electors would be able to vote formally on any proposals for reform. At the time of writing, the government is proposing three choices on the ballot:

https://en.wikipedia.org/wiki/Popular_vote

https://en.wikipedia.org/wiki/Majority



• First-past-the-post, the current voting system;• Some form of proportional representation;• Preferential ballot, or ranked ballot.

The Special Committee of the Legislature has promised to craft a plebi-scite question in 2016 (Prince Edward Island, 2015, November 27).

Ontario, 2007

The Ontario Liberal Party had campaigned in 2003 on a program that in-cluded a range of electoral reforms. In November 2004, a year after being elected, Ontario Premier Dalton McGuinty, leader of the Liberal Party, is-sued a statement following a “Dialogue on Democracy” conference that said: “When it comes to how the people elect their representatives, the people of Ontario will have their say” (Ontario, 2004a, November 18).

Michael Bryant, the Attorney General and Minister Responsible for Democratic Renewal, stated that “We will involve Ontarians directly in improving the quality of our democracy, modernizing our political institu-tions, and restoring public faith in government” (Ontario, 2004).

The government promised to create a Citizens’ Assembly on Electoral Reform that would be free to consider all options. “It may be that Ontar-ians choose to keep our first-past-the-post system,” the premier indicated,

“That’s fine. The very exercise of re-examining our electoral system will re-invigorate and heighten our appreciation of it.” Significantly, he pointed out that “This is a matter for Ontarians to decide. Our responsibility is to ensure the public’s voice is heard loud and clear, and has an impact” (Mc-Guinty, 2004).

The premier reemphasized the point in Parliament when he declared that “We’re going to the citizens of Ontario. We believe the issue of elec-toral reform is so fundamental, so basic, that we’re asking the people of Ontario for their judgment in this matter” (McGuinty, 2004).

Kuldip Kular, the parliamentary assistant to the attorney general, de-clared that “Ontario’s electoral system belongs to Ontarians, not to elected



officials or appointed commissions. So we are asking Ontarians to decide for themselves how our political system should work and how they want to elect MPPs here to Queen’s Park. No government in this province has ever given citizens this kind of opportunity. This bill, if passed, will give the people of Ontario the chance to have their say on the role of money in politics and electoral reform” (Ontario, 2005, June 13).

The government decided to apply the same formula as in British Co-lumbia and PEI. For electoral change to be triggered, the alternative system would require 60% of voter support, as well as at least 50% support in 64 of the 107 (60% of total) ridings. The referendum in Ontario was held at the same time as the provincial election of October 2007. The Citizens Assem-bly had come to the conclusion that a mixed-member proportional system be adopted. The question put to the people was:

Which electoral system should Ontario use to elect members to the provincial legislature? • The existing electoral system (first-past-the-post)• The alternative electoral system proposed by the Citizens’ Assembly

(Mixed Member Proportional)

The proposal was defeated, with 36.9% of the electorate supporting the proposal, and 63.1 voting in favour of the existing electoral sys-tem. Five ridings (out of 107), all in central Toronto, gave a majority of support to the Citizens’ Assembly proposal. The turnout for the 2007 election was 52.8%, indicating that less than 20% of eligible voters sup-ported the proposal.

The convention of voting practice in Westminster systems

The Constitution of Canada was drafted with a commitment to live by the “principles” of the Westminster system. It is worthwhile to examine how other jurisdictions in the Westminster family have consulted the public to legitimize electoral reform.



The United KingdomFollowing the 2010 general election in the United Kingdom, the Conser-vative Party led by David Cameron and the Liberal Democratic Party led by Nick Clegg agreed on a coalition agreement which committed the gov-ernment to holding a plebiscite: “You will get a referendum on the voting system, so you have a greater say on who represents you in Parliament. Government will be transparent. You will be able to get your hands on all the information you need,” Deputy Prime Minister Nick Clegg declared (United Kingdom, Office of the Prime Minister, 2010, May 2010). That September, Clegg said to the House of Commons that “When a big ques-tion mark hangs over something as important as our voting system, the only way to resolve the dilemma is to let people have their say.” He em-phasized that “that the final decision should be made not by us, but by the British people” (Stevenson, 2010).

Prime Minister David Cameron emphasized the need for a clear public mandate. In January 2011, the prime minister said, through his spokesman, that a referendum was necessary in order to “allow the people to decide on voting reform and that a referendum was a demo-cratic step” (Cameron, 2011a). A month later, the prime minister de-clared, on behalf of Mr. Clegg and himself, that “Far above our beliefs about how the voting system should work, we share a much more im-portant belief—a belief in democracy and the voice of the people being heard”(Cameron, 2011b).

The referendum was held in May 2011. Eligible voters were asked to express either a “yes” or a “no” to the following question:

At present, the UK uses the “first past the post” system to elect MPs to the House of Commons. Should the “alternative vote” system be used instead?

With a turnout of 42.2%, the “No” won this campaign, registering 67.9% of the vote. 32.1% (or 13.5% of eligible voters) thus expressed their support for the alternative voting method.



New ZealandThe issue of electoral reform in New Zealand brewed for many years before changes were actually enacted. In all cases, the New Zealand government sought the support of the people. In 1984, the Labour Party committed itself to a reform agenda and established a Royal Commission once it was elected. That commission’s report recommended major changes; it also insisted that a referendum be held on the issue. New Zealand turned to voters three times in order to decide on electoral reform.

The first referendum, held in 1992, was technically non-binding and asked two questions. The first was whether the first-past-the-post system should be replaced or retained. The second asked voters to choose between four systems: Single Transferable Vote (STV) (a form of ranked ballots that is used to select a number of members of Parliament in a single riding), Alternative Vote (AV) (another variant of ranked voting), Supplementary Member (SM) system (a preferred ballot system, but one where only two choices can be made), or Mixed Member Proportional (MMP) (a two-tier system involving the election of a local member of Parliament based on traditional voting methods, but also members of Parliament drawn from a list and who are elected according to the proportion of the vote the party received in the election). New Zealanders voted 84.7% in favour of change, with over 70% favouring the MMP model.

A second referendum was held in 1993 and, this time, it was binding. It asked voters to choose either the Mixed Member Proportional method or the traditional system: 53.86% of voters supported the MMP proposal and 46.14% supported first-past-the-post.

Based on majority support, New Zealand immediately introduced MMP. The experience was mixed, and in 2008 the National Party proposed another referendum. Prime Minister John Key declared that “Finally, we’ll open our ears to New Zealanders’ views on their voting system” (Key, 2008). This referendum, like the one held in 1992, was to be non-binding. It gave voters the choice of either maintaining the MMP system, or to change it. The second question asked voters to choose among first-past-the-post, Alternative Vote (or ranked voting), Single Transferable Vote, or Supple-mentary Member schemes.



New Zealanders voted to keep the MMP system (56.17%). Of those who wanted to scrap the MMP system, 31.19% of voters chose the first-past-the-post system, with the other three systems earning much less popularity.

Australia: The Australian Capital Territory (1992)Australia has long made important changes to its electoral system without consulting the public. That changed when a referendum was held in the Australian Capital Territory (ACT) in February 1992. The ACT was, at the time, a relatively small jurisdiction of about 300,000 people, which tended to produce dramatically lopsided electoral results.

The 1992 referendum was an “advisory poll” held simultaneously with the ACT’s Assembly election. The question simply asked if voters preferred the traditional “first-past-the-post” method or the Single Transferable Vote (or ranked vote) scheme. With a turnout of over 90%, voters chose the STV system over the single-member electoral system by a margin of 65.3% to 34.7%.

It is clear that other Westminster systems have also considered elec-toral change. What is remarkable is that in the last 25 years, governments felt compelled to allow the voters to have a say. The Canadian practice at the provincial level was thus consistent with other systems that have oper-ated under the “principles of the United Kingdom.”

Less important issues have been put to the people

Electoral reform touches the fibre of a political culture, so it is not surpris-ing that Westminster jurisdictions that have contemplated change have referred the question to the electorate. Indeed, on many occasions in his-tory, Canadians were asked to express themselves on far less consequential matters.

In September 1898, for instance, a national referendum was held on whether a prohibition of alcohol should be implemented. Twenty-one



years later, a referendum was held in Quebec asking voters if they wished to see the sale of alcohol legalized again. In 1988, Prince Edward Island held a plebiscite to approve the idea of a bridge that would link the island to the mainland. In Newfoundland and Labrador, voters were asked twice to answer specific policy choices for their educational system: In Septem-ber 1995, a referendum was held on the abolition of denominational edu-cation system. Two years later, another referendum was held on the estab-lishment of non-denominational education system.

In 1997, Alberta held a series of referenda at the local level on the in-stallation of Video Lottery Terminals. Prince Edward Island did the same the following year, as did New Brunswick in 2001. In October 2004, Nova Scotia voters were asked to express themselves on Sunday shopping. Brit-ish Columbia held a referendum on retaining the Harmonized Sales Tax in the summer of 2011.

Some broader policy issues have also been put to the people. In April 1942, the Canadian government asked voters to release it from a promise not to implement conscription. In 2002, British Columbia held a referen-dum on treaty negotiations. According to Premier Gordon Campbell, the motivation to go to the people was to foster engagement: “We believe it’s time the public was included in this process,” he said. “The referendum will give British Columbians a direct say on the principles that we believe should guide the province’s approach to treaty negotiations” (Campbell, 2002, April 27).

Matters of critical importance to the governance of Canada have also been the subject of referenda. Citizens of Newfoundland and Labrador were given an opportunity in June and July 1948 to say if they wanted to become a part of Canada or remain independent. Quebeckers were given two opportunities to express themselves on their national destiny, in May 1980 and again in October 1995. In August 1992, a national referendum was held to approve the Charlottetown Accord.

In sum, Canadians have been asked to vote on particular issues and most were less consequential than electoral reform. In asking Canadians for their input, important precedents were set.



Conclusion

The Jennings test on the validity of the conventional rule can thus be applied to the necessity of seeking popular agreement on electoral reform. To the Jennings question of “What are the precedents?” this chapter has provided the record from Canada and abroad over the past two decades. In all cases, governments turned to referenda to seek the approval of voters.

The clear evidence is that, over the past generation, both in Canada and abroad, electoral reform proposals have been put to people. Governments big and small felt compelled by the idea that no changes to the electoral system could implemented without the expressed consent of a majority (sometimes a supermajority) of the electorate. National governments in the United Kingdom and New Zealand have done so. So have governments in large Canadian provinces such as Ontario and British Columbia, and relatively small ones such as Prince Edward Island. Governments domi-nated by Progressive Conservatives felt the need to consult the electorate, as did Liberal and Labour governments. Plebiscites were used by govern-ments in minority or coalition situations as well as by governments in ut-terly dominating positions. It has been, quite simply, the established prac-tice—the very definition of convention—to consult voters about changes to the way representatives are elected.

To the Jennings question of “Did the actors in the precedents believe that they were bound by a rule?” the answer is equally clear. All the leaders felt compelled to make the argument that they needed a mandate from the people in order to proceed with whatever reform was being proposed by their parliamentary or extra-parliamentary commissions.

In each case, the government leaders felt compelled to give the reasons why. First, there was a recognition that no government—even one elected with a crushing electoral victory—had the mandate to proceed unilaterally. Second, the matter was judged to be intrinsically of such a fundamental nature that it was inconceivable to think that a transformation of vote-counting practices could be implemented without a solid expression in favour of it by the people. Not least, even governments that had included



electoral reform in their election platforms felt obliged to seek specific approval for those changes from voters.

In all cases, the governments took care to ask clear, concise questions that would yield either a mandate to continue or a warning to stop. Simi-larly, the various governments sponsored commissions and consultations to educate the electorate so as to engage it in considering the basis of the electoral system.

To the final Jennings question of whether there exists “a constitutional reason for the rule,” the answer lies in the preamble of the Canadian Con-stitution. Canada adopted a Westminster system of Parliament that in turn created a balance of power between the crown, the Houses of Parliament, and the Courts. The electoral system was a fundamental part of that bar-gain, based on conventions. It follows that any change to that equilibrium would be constitutional. Jennings’ questions thus must be answered in the affirmative in this case.

The Canadian electoral system has functioned on a system of conven-tions—understandings based on precedents, a recognition that going to the people was imperative, and that the issue was just as significant, if not more so, than other questions that have been put to the people. This real-ity was recognized by all the provinces that had made promises in their electoral platforms to consider reforms in how Canadians would be repre-sented in legislative assemblies. In all cases, these governments considered it crucial to refer the question to the people. Why should it be different for the government of Canada? There is now a moral imperative to put the issue to the people. As Peter Hogg put it in his authoritative book Consti-tutional Law of Canada, “there is a stronger moral obligation to follow a convention than a usage, and that departure from convention may be criti-cized more severely than departure from usage” (Hogg, 2005: 24).

The way we vote shapes our political culture. Canada is not perfect, and its democracy has its faults. But it must be recognized that the sys-tem has worked and the electorate that has sanctioned the system for generations needs to be consulted. The government of Canada simply cannot assume that it can unilaterally change the way in which we vote. It has no exclusive claim to the electoral process and it must respect con-



ventions. The precedents set in Canada and in other Westminster systems over the past 20 years dictate this necessity.

The fact that electoral reform has already been rejected four times by Canadians in plebiscites makes the matter all the more imperative. The past views that voters have expressed cannot simply be discarded. As is the case in any other jurisdiction, the federal government must conduct rig-orous and comprehensive consultations that are not simply driven by the self-appointed advocates of reform. Beyond that, the process must include a referendum, no matter how much it costs or how long it delays decisions. Regardless of the result, the government must abide by it. Without going to the people, it can expect no legitimacy to make any changes to the pre-cious process of elections, the essential tool of our democratic civilization.

References

British Columbia (2003, April 28). Nos. 52 and 53. Votes and Proceedings of the Legislative Assembly of British Columbia. Legislative Assembly of British Columbia. <https://www.leg.bc.ca/Pages/BCLASS-Legacy.aspx#%2Fcontent%2Flegacy%2Fweb%2F37th4th%2Fvotes%2Fv030428.htm>, as of June 6, 2016.

British Columbia (2003, April 30). Official Report of the Debates of the Legislative Assembly: 4th Session, 37th Parliament, Afternoon Sitting. Hansard 14, 12. Legislative Assembly of British Columbia. On-line version: <https://www.leg.bc.ca/documents-data/debate-transcripts/37th-parliament/4th-session/h30430p> and print version: <https://www.leg.bc.ca/content/hansard/37th4th/H0430pm-12.pdf>, as of June 6, 2016.

British Columbia, Citizens’ Assembly on Electoral Reform (2004). Making Every Vote Count: The Case For Electoral Reform In British Columbia. Final Report (December). Citizens’ Assembly on Electoral Reform. <http://participedia.net/sites/default/files/case-files/653_265_final_report.pdf>, as of June 6, 2016.

https://www.leg.bc.ca/Pages/BCLASS-Legacy.aspx#%2Fcontent%2Flegacy%2Fweb%2F37th4th%2Fvotes%2Fv030428.htm

https://www.leg.bc.ca/Pages/BCLASS-Legacy.aspx#%2Fcontent%2Flegacy%2Fweb%2F37th4th%2Fvotes%2Fv030428.htm

https://www.leg.bc.ca/documents-data/debate-transcripts/37th-parliament/4th-session/h30430p

https://www.leg.bc.ca/documents-data/debate-transcripts/37th-parliament/4th-session/h30430p

https://www.leg.bc.ca/content/hansard/37th4th/H0430pm-12.pdf

https://www.leg.bc.ca/content/hansard/37th4th/H0430pm-12.pdf

http://participedia.net/sites/default/files/case-files/653_265_final_report.pdf

http://participedia.net/sites/default/files/case-files/653_265_final_report.pdf



British Columbia (2005, September 12). Speech from the Throne. Opening Session, 38th Parliament. British Columbia. <https://www.leg.bc.ca/content/legacy/web/38th1st/Throne_Speech_2005_1st_38th.pdf>, as of June 6, 2016.

British Columbia Liberal Party (2001). A New Era for British Columbia: A Vision for Hope and Prosperity for the Next Decade and Beyond. BC Liberal Party. <https://www.poltext.org/sites/poltext.org/files/plateformes/bc2001lib_plt._27122008_141728.pdf>, as of June 6, 2016.

Cameron, David (2011a, January 18). Number 10 Press Briefing – Morning. Press Release. Government of the United Kingdom. <https://www.gov.uk/government/news/number-10-press-briefing-morning-from-18-january-2011>, as of June 14, 2016.

Cameron, David (2011b, February 18). Votes Referendum: Cameron Speech in Full. BBC News. <http://www.bbc.com/news/uk-politics-12504935>, as of June 14, 2016.

Campbell, Gordon (2002, April 27). Referendum Will Spur Treaty Process. National Post: A22. <http://www.vcn.bc.ca/~dastow/npo20427.txt>, as of June 6, 2016.

Hogg, Peter (2005). Constitutional Law of Canada, 2005 Student Edition. Thomson Carswell.

Jennings, Ivor (1959). The Law and the Constitution, 3rd edition. University of London Press.

Key, John (2008, August 3). John Key Speech to Annual National Conference. Scoop. <http://www.scoop.co.nz/stories/PA0808/S00023.htm>, as of June 9, 2016.

McGuinty, Dalton (2004). Remarks by Dalton McGuinty, Premier of Ontario, at the Dialogue On Democracy Conference. Archived Bulletin (November 18).

https://www.leg.bc.ca/content/legacy/web/38th1st/Throne_Speech_2005_1st_38th.pdf

https://www.leg.bc.ca/content/legacy/web/38th1st/Throne_Speech_2005_1st_38th.pdf

https://www.poltext.org/sites/poltext.org/files/plateformes/bc2001lib_plt._27122008_141728.pdf

https://www.poltext.org/sites/poltext.org/files/plateformes/bc2001lib_plt._27122008_141728.pdf

https://www.gov.uk/government/news/number-10-press-briefing-morning-from-18-january-2011



http://www.bbc.com/news/uk-politics-12504935

http://www.vcn.bc.ca/~dastow/npo20427.txt

http://www.scoop.co.nz/stories/PA0808/S00023.htm



Government of Ontario. <https://news.ontario.ca/opo/en/2004/11/remarks-by-dalton-mcguinty-premier-of-ontario-at-the-dialogue-on-democracy-conference.html>, as of June 14, 2016.

Ontario (2004a, November 18). Ontarians to have a Say on Electoral Reform. News release. Government of Ontario. <https://news.ontario.ca/archive/en/2004/11/18/Ontarians-to-have-a-say-on-electoral-reform.html>, as of June 6, 2016.

Ontario (2004b, November 18). Statements by the Ministry and Responses: Democratic Renewal/Renouveau Démocratique. Hansard Transcripts: Official Records for 18 November 2004. Legislative Assembly of Ontario. <http://www.ontla.on.ca/web/house-proceedings/house_detail.do?Date=2004-11-18&Parl=38&Sess=1&locale=en#P390_124300>, as of June 6, 2016.

Ontario (2005, June 13). Election Statute Law Amendment Act, 2005. Hansard Transcripts: Official Records for 13 June 2005. Legislative Assembly of Ontario. <http://www.ontla.on.ca/web/house-proceedings/house_detail.do?Date=2005-06-13&Parl=38&Sess=1&locale=en#P867_134251>, as of June 6, 2016.

Prince Edward Island (2003, January 21). Commissioner Appointed to Consider Electoral Reform. News release. Government of Prince Edward Island. <http://www.gov.pe.ca/newsroom/index.php?number=news&dept=&newsnumber=2913>, as of June 1, 2016.

Prince Edward Island Electoral Reform Commission (2003). 2003 Prince Edward Island Electoral Reform Commission Report. PEI Electoral Reform Commission. <http://www.gov.pe.ca/photos/original/er_premier2003.pdf?PHPSESSID=39e691d%203388c41a312d95cfd75a7a6bf>, as of June 1, 2016.

Prince Edward Island (2004, December 16). Oral Questions. Second Session, 62nd General Assembly. Legislative Assembly of Prince Edward Island.

https://news.ontario.ca/opo/en/2004/11/remarks-by-dalton-mcguinty-premier-of-ontario-at-the-dialogue-on-democracy-conference.html



https://news.ontario.ca/archive/en/2004/11/18/Ontarians-to-have-a-say-on-electoral-reform.html

https://news.ontario.ca/archive/en/2004/11/18/Ontarians-to-have-a-say-on-electoral-reform.html

http://www.ontla.on.ca/web/house-proceedings/house_detail.do?Date=2004-11-18&Parl=38&Sess=1&locale=en#P390_124300




http://www.gov.pe.ca/newsroom/index.php?number=news&dept=&newsnumber=2913


http://www.gov.pe.ca/photos/original/er_premier2003.pdf?PHPSESSID=39e691d 3388c41a312d95cfd75a7a6bf

http://www.gov.pe.ca/photos/original/er_premier2003.pdf?PHPSESSID=39e691d 3388c41a312d95cfd75a7a6bf



<http://www.assembly.pe.ca/sittings/2004fall/hansard/2004-12-16-hansard.pdf>, as of June 14, 2016.

Prince Edward Island (2005, February 17). Appointments Announced to the Commission on Prince Edward Island’s Electoral Future. News release. Government of Prince Edward Island. <http://www.gov.pe.ca/newsroom/index.php?number=news&dept=&newsnumber=4026>, as of June 6, 2016.

Prince Edward Island (2005, November 16). Speech from the Throne, 3rd Session, 62nd General Assembly: A Clear Direction... Putting Islanders First. Legislative Assembly of Prince Edward Island. <http://www.assembly.pe.ca/speech/2005/index.php>, as of June 6, 2016.

Prince Edward Island (2015, November 27). First Session, 65th General Assembly. Hansard. Legislative Assembly of Prince Edward Island. <http://www.assembly.pe.ca/sittings/2015fall/hansard/2015-11-27-hansard.pdf>, as of June 6, 2016.

SCC Re: Resolution to amend the Constitution, [1981] 1 S.C.R. 753 at 888–909. “Patriation Reference.” <http://scc-csc.lexum.com/scc-csc/scc-csc/en/2519/1/document.do>, as of June 9, 2016.

Stevenson, Alex (2010, September 7). Electoral Reform Rebels Hold Fire – For Now. politics.co.uk. <http://www.politics.co.uk/news/2010/09/07/electoral-reform-rebels-hold-fire-for-now>, as of June 14, 2016.

United Kingdom, Office of the Prime Minister (2010, May 20). The Coalition, Our Programme for Government. Press conference. Government of the United Kingdom. <https://www.gov.uk/government/speeches/the-coalition-our-programme-for-government-press-conference>, as of June 6, 2016.

http://www.assembly.pe.ca/sittings/2004fall/hansard/2004-12-16-hansard.pdf




http://www.assembly.pe.ca/speech/2005/index.php

http://www.assembly.pe.ca/speech/2005/index.php



http://scc-csc.lexum.com/scc-csc/scc-csc/en/2519/1/document.do

http://scc-csc.lexum.com/scc-csc/scc-csc/en/2519/1/document.do

http://www.politics.co.uk/news/2010/09/07/electoral-reform-rebels-hold-fire-for-now

http://www.politics.co.uk/news/2010/09/07/electoral-reform-rebels-hold-fire-for-now

https://www.gov.uk/government/speeches/the-coalition-our-programme-for-government-press-conference

https://www.gov.uk/government/speeches/the-coalition-our-programme-for-government-press-conference



Publishing information

Distribution

These publications are available from <http://www.fraserinstitute.org> in

Portable Document Format (PDF) and can be read with Adobe Acrobat® or Adobe

Reader®, versions 7 or later. Adobe Reader® XI, the most recent version, is available

free of charge from Adobe Systems Inc. at <http://get.adobe.com/reader/>.

Readers having trouble viewing or printing our PDF files using applications from

other manufacturers (e.g., Apple’s Preview) should use Reader® or Acrobat®.

Media

For media enquiries, please contact our communications department at

604.714.4582 or via e-mail at [email protected]. In

Toronto, contact our media specialist at 416.363.6575, ext. 238.

Supporting the Fraser InstituteTo learn how to support the Fraser Institute, please contact

•Development Department, Fraser Institute, Fourth Floor, 1770 Burrard Street,

Vancouver, British Columbia, V6J 3G7 Canada

• telephone, toll-free: 1.800.665.3558 ext. 586

• e-mail: [email protected]

•website: <http://www.fraserinstitute.org/support-us/overview.aspx>

AcknowledgmentsThe authors wish to thank the Lotte and John Hecht Memorial Foundation for their

generous funding of this project.



Purpose, funding, and independenceThe Fraser Institute provides a useful public service. We report objective informa-

tion about the economic and social effects of current public policies, and we offer

evidence-based research and education about policy options that can improve the

quality of life.

The Institute is a non-profit organization. Our activities are funded by charit-

able donations, unrestricted grants, ticket sales, and sponsorships from events, the

licensing of products for public distribution, and the sale of publications.

All research is subject to rigorous review by external experts, and is conducted

and published separately from the Institute’s Board of Trustees and its donors.

The opinions expressed by authors are their own, and do not necessarily reflect

those of the Institute, its Board of Trustees, its donors and supporters, or its staff.

This publication in no way implies that the Fraser Institute, its trustees, or staff are

in favour of, or oppose the passage of, any bill; or that they support or oppose any

particular political party or candidate.

As a healthy part of public discussion among fellow citizens who desire

to improve the lives of people through better public policy, the Institute

welcomes evidence-focused scrutiny of the research we publish, includ-

ing verification of data sources, replication of analytical methods, and

intelligent debate about the practical effects of policy recommendations.

About the Fraser InstituteOur mission is to improve the quality of life for Canadians, their families, and

future generations by studying, measuring, and broadly communicating the

effects of government policies, entrepreneurship, and choice on their well-being.

Notre mission consiste à améliorer la qualité de vie des Canadiens et des générations

à venir en étudiant, en mesurant et en diffusant les effets des poli tiques gouverne-

mentales, de l’entrepreneuriat et des choix sur leur bien-être.



Members

Past members

Editorial Advisory Board

* deceased; † Nobel Laureate

Prof. Terry L. Anderson

Prof. Robert Barro

Prof. Michael Bliss

Prof. Jean-Pierre Centi

Prof. John Chant

Prof. Bev Dahlby

Prof. Erwin Diewert

Prof. Stephen Easton

Prof. J.C. Herbert Emery

Prof. Jack L. Granatstein

Prof. Herbert G. Grubel

Prof. James Gwartney

Prof. Ronald W. Jones

Dr. Jerry Jordan

Prof. Ross McKitrick

Prof. Michael Parkin

Prof. Friedrich Schneider

Prof. Lawrence B. Smith

Dr. Vito Tanzi

Prof. Armen Alchian*

Prof. James M. Buchanan* †

Prof. Friedrich A. Hayek* †

Prof. H.G. Johnson*

Prof. F.G. Pennance*

Prof. George Stigler* †

Sir Alan Walters*

Prof. Edwin G. West*

copyright ©2016 by the fraser institute. all rights ... · on june 7, 2016, the government formed...

Documents